This study proposes a novel analysis system for printed fabrics that can automatically make color separation and identify repeat patterns. The system uses a scanner to obtain red, green and blue (RGB) color images of printed fabrics and then convert them into hue, saturation, intensity (HSI) color images. In order to obtain color separation, a genetic algorithm is used to search for a smaller sub-image with the same color distribution, and then the color separation is conducted by use of the recursive region splitting method. Then carry out another Fuzzy C-means (FCM) calculation on the HSI image using the color clusters (cluster number) and values (cluster centers) obtained from separating the colors of sub-images to quickly classify colors for the pixels. Pixels of different color categories are marked with different gray levels. In this way, a polychromatic pattern image is formed. For identifying repeat patterns, first, a template matching method is applied to discover distributions of same pattern elements. Then, the Hough transform method is used to obtain the cutting positions and dimensions of the repeat patterns in the polychromatic pattern image. Next, the images of the repeat patterns are extracted out from the polychromatic images. Finally, the repeat units of the black pictures are generated based on the color categories and they are expanded to become black pictures that can be used to make plates. According to the experimental results, this system can rapidly and automatically separate colors and identify repeat patterns of images on printed fabrics.