Objective According to the theory of traditional Chinese medicine(TCM)，the human head and face colors can reflect the health status of a person. As the traditional method of judging the color of zygomatic depending on the doctor’s visual judgment，the paper proposes a method based on support vector machine(SVM)that can classify zygomatic color automatically. Methods Firstly，many zygomatic red and non-zygomatic red 8×8 blocks are selected by TCM experts in original RGB color space. Then，we extract the R，G，B values from the 64 pixel block as the features，train the features(extracted from that blocks)and classify each block of zygomatic images as red or not red，and the ratio of the red blocks in the two zygomatic regions can be computed. Finally，we can determine a person belonging to zygomatic red or not according to that ratio. Results The experiments are implemented in the image library which is identified by TCM experts and the experimental results show that the classification accuracy is close to 83%. Conclusions The classification effect of zygomatic color based on support vector machine method is valid.