With the progress of the times, the aesthetics of the public have gradually become diversified, and clothing art design should also keep pace with the times, and carry out the innovation of clothing art design under the integration of traditional graphics. In this paper, the RGB and HSV values of traditional graphics are extracted by the method of color feature extraction, and each pixel is detected by scanning pixel by pixel, and the color characteristics of the graphics are judged according to the pixel points. The roughness and sharpness of the graphic texture are characterized by the gray symbiotic texture features. Combined with the central moments reflecting the distribution of gray levels in the object area relative to the gray level center of gravity metric, the size, shape, and other details of the image are described with the help of higher-order moments of more than three orders. The main direction of clothing art design innovation is analyzed by integrating different features of traditional images into clothing pattern design. The results show that the RGB values in the color design of sweet style clothing should be similar to (255,192,203), (135,206,250), (0,51,0), the entropy value of the graphic texture in the design of clothing art should be >1, and the degree of regularity should be >0.9, so that the regularity and regularity of the graphic is reflected as much as possible. The shape characterization value Φ6 and Φ7 of the image in clothing art design should be 0 to prevent the deformation caused by image scaling and rotation.