With the advent of the digital age,Content-based image retrieval has become a research hotspotand image feature extraction is an important research field of content-based image retrieval. Uygur herbs is fundamental guarantee of preventing and curing disease of xinjiang uygur medicine.Uygur herbs play an important role for the health of the people in xinjiang.According to the difference of Uygur herbs application site,it can be divided into flower herbs and leafy herbs.First,Preprocessing,including imagecolor enhancement,size normalizition and color space transformation.Then extracting color and textural features as acomprehensive feature and analysing them with statistics method.At last,the classification ability of featuresis evaluated by Bayes discriminant analysis.The experimental results show that the image classification accuracy of Uygur herbs by using the comprehensive feature is higher than the single use of color or textural feature.This research laid a certain foundation for the content-based retrieval research of xinjiang uygur herbs image.