Photoshop Generative Fill: How to Combine Two Images

Updated 8 November 2026 · Standard formula, shown below

To combine two images with Photoshop Generative Fill, place one image over the other, select the area where they should meet, and use Generative Fill to create or extend pixels that blend the boundary. For a reliable composite, first align and mask the images yourself; Generative Fill is best used to fill gaps and soften transitions, not to place an exact second photograph automatically.

How Generative Fill fits into a composite

Generative Fill is a selection-based editing feature in Photoshop. It can generate content inside a selected area or replace selected pixels, using the surrounding image and an optional text prompt as guidance. It does not simply merge two files or guarantee that a generated result will preserve every detail of a source image.

A practical composite has three parts: a background image, a second image on its own layer, and a mask or selection defining which parts of that layer remain visible. Once the two sources are positioned, Generative Fill can help create missing background, extend a sky or wall, or make a transition less abrupt.

For example, if a person photographed outdoors needs to appear in a different landscape, place the person’s image above the landscape, remove or mask the original background, then use Generative Fill to repair small gaps around the silhouette or add missing landscape around the edges. Keep the person’s original pixels wherever likeness, clothing, or fine details matter.

Prepare and place the two images

Open the intended background image in Photoshop.
Use File > Place Embedded to add the second image as a separate layer. Placing it this way keeps it independently transformable and avoids flattening the sources together.
Use the Move tool and transform controls to position and scale the upper layer. Hold the proportions when resizing if needed, and check that the subject’s perspective, size, and light direction make sense against the background.
Remove the unwanted background from the upper image with a layer mask or a selection tool. A mask is generally safer than erasing because you can restore hidden areas by painting on the mask.
Save a layered working file before generating content. This preserves the originals and makes it easier to revise the composite.

Match the images before blending. A sharply lit subject against a soft, overcast background can still look pasted in even if the edge is clean. Compare horizon height, camera angle, focus, color temperature, and shadow direction. Generative Fill may invent plausible pixels, but it cannot make inconsistent source photography physically consistent by itself.

Use Generative Fill to repair the join

Select the layer that should receive the generated pixels. The active layer matters: Photoshop may not fill the area you intend if a different layer is selected.
Use a selection tool, such as the Lasso or Object Selection tool, to outline a gap or the area needing a transition. For a boundary repair, include a modest strip of surrounding pixels so the generated result has context. Avoid selecting across details you must preserve.
Choose Generative Fill in the contextual task bar or use the available Generative Fill command. If Photoshop requests a prompt, describe only what should appear in the selected region, such as “continue the overcast sky” or “soft grass beside the path.” Leaving the prompt empty can be useful when you want Photoshop to infer a continuation from the image.
Generate the result, then inspect the available variations in the Properties panel. Choose the closest option, or adjust the selection or prompt and generate again.
Use a layer mask, a soft brush, or a smaller follow-up selection to refine the transition. Keep generated content away from important subject features unless replacing those features is intentional.

Selections are not masks: a selection temporarily specifies where an operation applies, while a layer mask controls visibility on a layer. For a nondestructive composite, use a mask to blend the placed image and use Generative Fill on a separate layer when your Photoshop workflow allows it. This makes it easier to compare the generated result with the untouched sources.

Write prompts and selections that stay focused

Generative Fill usually works better when the selected area has enough neighboring visual information and the instruction is limited to the task. If a narrow seam needs matching, select the seam plus a little context and ask for a continuation of the nearby material. A long prompt describing the whole scene can distract from the local repair.

For a missing edge: select beyond the canvas edge and ask for a continuation of the visible setting.
For a cutout fringe: make a small selection around the fringe and describe the background material that should show through.
For a hard transition: blend the upper layer with its mask first, then generate only where the two images leave holes or mismatched areas.
For an exact object: place and mask the object from the source photograph instead of asking Generative Fill to recreate it. Generated details may differ from the original.

Do not expect a prompt to preserve precise text, logos, facial details, or product features. Generated content can change or invent details. When accuracy matters, retain the original source pixels and confine generation to surrounding areas.

Common problems and fixes

The generated area looks unrelated: reduce the selection, include more relevant nearby context, and simplify the prompt. Try another variation rather than repeatedly changing several things at once.
A halo appears around the cutout: refine the layer mask and inspect the edge at high zoom. A halo can come from leftover color in the original background, not from the generated fill.
The composite looks pasted on: revisit scale, perspective, focus, color, and light direction. Use masks and ordinary tonal or color adjustments to match the source images; generation is not a substitute for those corrections.
The wrong layer changes: check the selected layer and selection before running the command. If possible, create a new empty layer for generated content and confirm the command is set to use the intended visible layers.
Generative Fill is unavailable: check that Photoshop is up to date, that the document and selected layer meet the feature’s requirements, and that the feature is available for your account and region. Availability and interface details can change, so consult Adobe’s current Photoshop help if the command is missing.

When to use another Photoshop method

Generative Fill is useful when the composite needs plausible new pixels, such as extending a background or repairing a small gap. It is less suitable for simple, exact compositing. For that, place the images as layers, use a layer mask, and make conventional adjustments. The subject remains the actual photographed subject rather than a generated approximation.

Content-Aware Fill can be a useful alternative for removing an object or filling a relatively uniform area from nearby image content. Clone Stamp and Healing tools provide more direct control for small repairs. These methods do not all solve the same problem: masking decides which source is visible, while fill tools synthesize or copy pixels into an area.

Generative features can require an internet connection or depend on account, region, and service availability. Check the current Photoshop documentation for eligibility and usage conditions. If the result is for a factual, editorial, or commercial context, follow the relevant disclosure and image-authenticity requirements; do not present generated scene details as documentary evidence.

Frequently Asked Questions

Can Photoshop Generative Fill combine two photos automatically?
It can generate or alter pixels in a selection, but it is not a one-click tool that reliably merges two complete photographs. Place the images on separate layers, align them, and mask the unwanted parts before using Generative Fill to repair or extend the surroundings.

Do I need to select both images?
No. Typically, you place both images in one document as layers, then select the specific gap or region that needs generated content. Choose the layer or setup that should receive the fill before running the command.

Should I leave the prompt blank?
You can try an empty prompt when the surrounding pixels make the intended continuation clear. For a specific material or setting, a short, direct description can help. Compare variations and refine the selection if the result is off target.

How do I blend the edges of the two images?
Use a layer mask to reveal or hide portions of the upper image, and refine the mask with a brush. Generative Fill can help repair gaps along the edge, but a mask gives direct control over which original pixels remain visible.

Why does Generative Fill change details I wanted to keep?
The selected region is eligible to be regenerated, and generated pixels are not guaranteed to match the source exactly. Make a smaller selection, exclude important details, or retain those details on the original layer or mask.

Method: the standard formula for this calculation, rounded as each result states. See how the calculators work or report an error.