10 Tips To Write Prompts In Google Nano Banana Pro For Creating That Perfect Picture

Google has given Nano Banana a massive upgrade. Built on Gemini 3, Nano Banana Pro (Gemini 3 Pro Image) is its most advanced image model to date. Nano Banana Pro includes advancements in text rendering, world knowledge, creative controls, and more. 

Before you get started with image creation, here are some tips to write prompts in Nano Banana Pro, as per Google.

1. Subject: Define what or who is the focus of the picture. For example, a woman in a spaceship with neon-pink hair, a white owl wearing goggles, etc.

2. Composition: Specify how is the frame arranged. For instance, close-up, aerial view, wide-angle shot, low-angle, portrait, etc.

3. Action: What’s taking place in the picture (monkey leaping across rooftops at dusk, artist painting a teacup, rocket soaring through storm clouds).

4. Location: Where is the scene set? For example, a rain-soaked Chicago alley, an enchanted forest at twilight, a Martian desert under twin moons).

5. Style: Lay out what visual language should the picture speak: hyper-realistic digital painting, vintage 35mm film grain, Studio Ghibli animation, etc.

6. Editing Instructions: When altering an existing image, be precise: replace the red jacket with matte black leather, erase the billboard on the left, add soft morning fog.

While basic prompts still produce good results, professional-level output demands richer direction. You can improve your prompts by including these details:

7. Composition And Aspect Ratio: Set the exact canvas (e.g., a square 1:1 album cover, a 21:9 cinematic frame).

8. Camera And Lighting: Direct the shot like a cinematographer. For example, “A low-angle shot with a shallow depth of field,” “Golden hour backlighting creating long shadows,” etc.

9. Text Integration: Spell out if any text should appear (e.g., “The words ‘Man of the Match’ on the giant screen”).

10. Reference Inputs: Nano Banana Pro can now integrate multiple input images to create a single finished output. When uploading multiple images, specify the role of each: Use Image A/1st for character, Image B/2nd for background, and Image C/3rd for props, etc.