Nano Banana Prompt: “Fix Just One Point, and Your Image Instantly Improves”
When generating images with Nano Banana, the quality is undeniably high, but often they don’t stick in your memory. I experienced the same thing. Trying to fix it by writing longer prompts only made things worse.
The conclusion was this. Don’t try to cram everything into one image; instead, set a single “Subject + Action” pair first. Once this is done, the image comes alive even if you remove the rest.
Today’s Core Key
Just one motif. Decide on these two things first: Subject (Who) + Action (What they are doing).
What is a Motif?
It is “the largest element in the frame that conveys the theme at a glance.” For example:
- Girl + Hugging a doll
- Wizard + The moment of casting a spell with a staff
- Knight + The moment of drawing a sword
- Student + The moment of taking a selfie with a phone
The important thing is not just the “Girl,” but fixing the “Action the girl is doing” together as one set. Characters just standing there tend to become boring.
Subject: 1 girl, single character only
Composition: 9:16, medium shot, character centered, large dominant motif is the shelter under the umbrella
Action: holding umbrella high, looking up at the rain, emphasize calm expression and hand gripping handle
Location: rainy night atmosphere, bokeh city lights in background, wet glass texture, low detail background
Style: anime key visual, cinematic lighting, moody, high quality
Focus rule: all visual interest concentrated on face and the rim of the umbrella, background softened into bokeh circles
Subject: 1 knight, single character only
Composition: 9:16, medium close-up, character placed diagonal, large dominant motif is the tension of drawing the sword
Action: hand gripping the sword hilt, pulling blade from scabbard, emphasize intense gaze and hand
Location: battlefield atmosphere, flying sparks and smoke, motion blur background, low detail
Style: cinematic fantasy art, dramatic lighting, high quality
Focus rule: all visual interest concentrated on the eyes and the sword hilt, strong depth of field, everything else simplified10 Ready-to-Use “Subject + Action” Candidates
Just pick one and start.
- Moment of brushing hair back with one hand and smiling
- Moment of saying a secret with a finger on lips (“Shh”)
- Moment of reaching out to pull someone in
- Moment of charging forward while drawing a sword
- Moment of tying a ribbon/string (Hands become the main focus)
- Moment of light bursting while unfolding a magic circle
- Moment of lifting an umbrella to block the rain
- Moment of pages blowing in the wind while opening a book
- Moment of focusing on music with earphones plugged in
- Moment of taking a picture with a camera
If you choose “Moment of charging forward while drawing a sword” as the motif, it looks like this.
Subject: 1 warrior, single character only
Composition: 9:16, dynamic low angle, dutch angle (tilted), large dominant motif is speed and impact
Action: body leaning forward aggressively, drawing sword from hip, emphasize dynamic motion and fierce gaze
Location: speed lines, heavy motion blur, flying dust particles, abstract battlefield background, low detail
Style: action anime style, high contrast, impact frame, high quality
Focus rule: all visual interest concentrated on the face and the blur of the sword being drawn, background turned into directional speed lines
Prompt: Single Dominant Motif Fixed Template (NanoBanana Pro)
Just change the parts in {} below. This version focuses on writing the core first, not writing at length.
Template
Subject: 1 {character_type}, single character only
Composition: {aspect_ratio}, {shot_type}, character placement is {placement}, large dominant motif is {dominant_motif}
Action: {one_clear_action}, emphasize {focus_area}
Location: {simple_background_mood}, minimal background, low detail
Style: {style_keywords}
Focus rule: all visual interest concentrated on {dominant_motif} and {focus_area}, everything else simplifiedExample (Ready to Use)
Subject: 1 female mage, single character only
Composition: 9:16 vertical poster, medium close-up, character centered slightly above midline, large dominant motif is "outstretched hand with glowing spell"
Action: casting a spell, emphasize face and outstretched hand
Location: simple dark gradient background with faint particles, minimal background, low detail
Style: anime key visual, clean lineart, soft rim light
Focus rule: all visual interest concentrated on the glowing spell and the hand, everything else simplified
One Failure Case + One Correction
1. Failure: Too Many Elements (Split Focus)
Feature: Mage, flashy spell, huge castle, and pet (dragon) are all described as “main characters,” scattering the gaze.
Prompt
Subject: 1 female mage, 1 large dragon pet
Composition: 16:9, wide shot, character and pet centered, large dominant motif is mage and dragon and castle
Action: casting complex explosion magic, interacting with dragon, holding detailed staff
Location: inside detailed fantasy castle, intricate architecture, stained glass windows, high detail background
Style: anime key visual, vibrant colors, high quality
Focus rule: high detail on face, dragon scales, magic effects, and castle bricks equally
Negative: low resolution, blur(Explanation)
- Subject: 1 female mage, 1 large dragon pet
- Composition: 9:16, wide shot, character and pet centered, large dominant motif is mage and dragon and castle (Everything emphasized)
- Action: casting complex explosion magic, interacting with dragon, holding detailed staff
- Location: inside detailed fantasy castle, intricate architecture, stained glass windows, high density background
- Focus rule: high detail on face, dragon scales, magic effects, and castle bricks equally (No resting place for the eyes)
Cause of Failure
- Put in 2 big motifs
- Example: “Mage+Spell” AND “Giant Castle” AND “Pet”…
- Result: The model draws everything, but the viewer asks, “So who is the protagonist?” This is a “No Focus” state.
2. Correction: Focus on One Large Motif (Selection and Concentration)
Feature: Keep only ‘the moment of using magic’, treat the castle as a silhouette, remove the pet, clean up props.
Prompt
Subject: 1 female mage, single character only
Composition: 9:16, medium shot, character placed center, large dominant motif is mage casting a spell
Action: casting glowing magic from hand, emphasize face and magic light
Location: dark night atmosphere, vague silhouette of a castle in distance, minimal background, low detail
Style: anime key visual, cinematic lighting, high quality
Focus rule: all visual interest concentrated on mage's face and glowing hand, background blurred and simplified(Explanation)
- Subject: 1 female mage, single character only (Pet removed)
- Composition: 9:16 (Vertical), medium shot, character centered, large dominant motif is mage casting a spell (Compressed into one)
- Action: casting glowing magic from hand, emphasize face and magic light
- Location: dark night atmosphere, vague silhouette of a castle in distance, minimal background, low detail (Castle exists only as atmosphere)
- Focus rule: all visual interest concentrated on mage’s face and glowing hand, background blurred and simplified
- Negative: extra characters, pet, animals, distracting background, props near face, complex architecture
Correction: Leave only one motif, treat the rest only as atmosphere, and leave the castle just as a “silhouette.”
Test
When you shrink the generated result on the screen, if you can read one thing at once, it’s a success.
- 1 thing: Face + Hand action is immediately visible -> OK
- 2~3 things: Face is visible, background is readable, props stand out -> Large motif failed
Conclusion
It’s easy to get “well-drawn images” with AI now, but “images that stop the eye” still require design. The start of design is not a grand composition, but fixing one motif (Subject + Action) first. Just doing this reduces blandness significantly.
