The work aims to extract characteristic colors of Mawangdui silk paintings, and analyze the color matching relationship between silk paintings, so as to provide guidance for the color matching process of cultural and creative product design and show the cultural characteristics of silk paintings more accurately in the design. The optimized K-means clustering algorithm was used to extract the characteristic colors of silk paintings, and a standard color card of silk paintings was generated. Then, a color network model was established with the color proportion data and color adjacency relationship in the image, and the matching of primary and secondary colors was determined. Finally, a number of color matching populations were automatically generated by CorelDRAW plug-in color shuffle, and the feasibility of extracting characteristic colors of silk paintings and the operability of assisting color matching with color network model were verified by the fuzzy comprehensive evaluation method. The standard color card of silk painting extracted by K-means clustering method is representative in color design. Building a color network model and plug-in assisted color matching can improve the color matching efficiency of designers and provide a new idea for the color matching design process of cultural and creative products.