The work aims to explore and present the color characteristics and matching relationship of bird patterns in Qiandongnan Miao Embroidery in a quantifiable and visual way, so as to provide color reference for the innovation of Miao pattern design, and prepare for the establishment of national color database. After 212 samples were denoised and smoothed with median filtering algorithm and adaptive texture filtering algorithm, K-Means clustering algorithm was used to extract image colors. The number of colors extracted was determined by automatically obtaining k values, and the color network model was constructed with dominant color first. With the help of the HSV color space model, the color features of different embroidery methods were summarized from hue, saturation and lightness by means of frequency statistics. The research results show that the color matching and regularity of bird patterns in Miao embroidery in the form of data and charts can objectively and accurately reflect the color characteristics of bird patterns in Miao embroidery. The research results can provide designers with a lot of reference for color matching schemes, improve the work efficiency of designers, and have important value for the redesign and application of national patterns and the inheritance and innovation of intangible cultural heritage.