Using Descriptive Words in a Stable Diffusion Prompt

Getting great results with stable diffusion starts with writing effective prompts. The prompt acts as the instructions for the AI to generate the image you want. Mastering the art of prompt engineering takes experimentation, but having some good examples to start with goes a long way.

In this article, I’ll share some of my favorite stable diffusion prompts I’ve used to generate stunning AI art, along with some tips on how to write your own.

Why Prompts Matter

Prompts may just look like a simple text description, but they allow you to tap into the full potential of stable diffusion. The more detailed and specific your prompt, the better the AI understands what you want to create.

Using the right keywords and structure in your prompt helps guide the AI, controlling aspects like:

  • Style and composition
  • Level of detail
  • Mood and lighting
  • Subject matter

Vague prompts lead to vague results. Dialing in an effective prompt takes experimentation, but the effort pays off in the quality of images produced.

Examples of Great Stable Diffusion Prompts

Here are a few of my favorite prompts that generate impressive images in stable diffusion:

Fantasy Landscape

An epic fantasy landscape digital painting, by Greg Rutkowski and Thomas Kinkade, trending on artstation, highly detailed

Fantasy landscape image

I love how this prompt generates a sweeping fantasy landscape with vivid colors and lighting. Specifying two artists helps define the style, while “trending on artstation” brings in elements that creative professionals currently find visually compelling.

Space Battle

A spaceship battle scene digital painting, by Syd Mead and Chris Foss, cinematic lighting, unreal engine, highly detailed

Space battle image

Chris Foss and Syd Mead are legends in sci-fi concept art, so I wanted their influence on an epic space battle scene. Adding “cinematic lighting” and “unreal engine” renders it with extra atmosphere and realism.

Magical Girl

Adorable magical girl with pink hair and blue eyes, wearing a frilly dress, by Hajime Sorayama and Disney, sparkles, soft lighting, intricate details

Magical girl image

For a cute anime-inspired magical girl, I blended Hajime Sorayama’s shiny futuristic style with Disney’s bubbly animation look. The extra descriptive details like hair color, sparkles, and lighting help define the style.

Tips for Writing Your Own Prompts

Here are some key tips to take your stable diffusion prompts to the next level:

Be Specific

Focus your prompt around a clear and specific subject, like “an astronaut walking on Mars” rather than just “space.” Give details on appearance, clothing, colors, etc.

Set the Mood

Describe the lighting, atmosphere and mood you want like “serene”, “cinematic”, or “mystical.” This helps set the tone.

Combine Artists

Name one or more artists whose styles you want to blend, like “by Greg Rutkowski and Alphonse Mucha.” This hybridizes their aesthetics.

Iterate and Improve

Don’t be afraid to try variations on a prompt and fine-tune over multiple generations. Pay attention to what works and build off of those iterations.

Helpful Resources

Here are some great sites to find prompts and inspiration from the stable diffusion community:

The key is to keep experimenting with new prompt ideas and find what works best for your creative vision!