Linos NEWS

How to Use Midjourney: Prompts, Parameters, and Image Creation

Step-by-step guide to creating images with Midjourney: prompt structure, parameters like aspect ratio and chaos, and tips for photorealistic or stylized results.

Linos NEWS Updated February 7, 2026 3 min read
Midjourney-style AI generated artwork
Midjourney-style AI generated artwork

How Midjourney Works

Midjourney is an AI image generator. You describe what you want in a text prompt; the model returns several image options. You can create from the web at midjourney.com on the Create page or via Discord. Images generate in real time on the Create page. You can also upload reference images to guide style or composition. Output quality depends on how you phrase the prompt and which parameters you use.

Writing Effective Prompts

A simple structure that works well is: subject + action or context + style or medium + lighting or color. Put the main subject first; words at the start of the prompt usually have more impact. Example: "A cybernetic samurai meditating under a neon waterfall, oil painting, golden hour." Be specific: "golden retriever in a coffee shop" is clearer than "dog in a room." For a particular look, name the style (e.g., 35mm photography, watercolor, vintage poster) or reference an artist or era (e.g., Art Nouveau, 1970s film).

Using Parameters

Parameters are extra instructions added at the end of the prompt. They use two dashes and no punctuation. Common ones include: --ar for aspect ratio (e.g., --ar 16:9 or --ar 2:3), --chaos for more variety among the four results (0–100), --quality for detail (often 0.25, 0.5, or 1), --stylize for how strongly to apply the Midjourney aesthetic, and --no to exclude elements (e.g., --no text, --no blur). Check the official Parameter List for your model version; syntax can change. Example: "vibrant California poppies --ar 2:3 --quality 1."

Getting Photorealistic Results

For photo-like images, use camera and lighting terms. Specify camera (e.g., Sony A7R IV, Nikon D850), lens (e.g., 35mm for street, 85mm for portrait), or film (e.g., Kodak Portra 400). Add lighting (e.g., golden hour, Rembrandt lighting, softbox) and environment (e.g., overcast, studio, backlit). Phrases like "shot on 35mm" or "documentary style" often improve realism. Avoid mixing with strong artistic styles in the same prompt.

Image-to-Image and References

On the Create page you can attach one or more images. Use them as style or composition references. The prompt then guides how the reference is used (e.g., "same style but with a different character"). This is useful for keeping a consistent look across several images or for basing a scene on a photo.

After Generation: Upscale and Variations

Midjourney returns a grid of options. You can upscale a chosen image for higher resolution or request variations (V1–V4) to get similar but different versions. Use the buttons or commands shown after the grid. Upscaling and some variations may use extra credits depending on your plan.

What to Do Next

Start with 5–10 prompts using the subject + context + style structure. Try different --ar values (e.g., 1:1, 16:9, 2:3) for different uses. For photos, add camera and lighting terms; for art, add medium and style. Check the official Midjourney docs for the latest parameters and Create page features.

Tags

midjourney ai art image generation how-to prompts

Related Articles