According to the characteristics of infrared thermal image,a pseudo-color coding method of infrared thermal image based on segmentation equalization was proposed to overcome the defects of the traditional pseudo color processing method.Infrared image is clustered by using improved K-means algorithm,and the various pixel’s coordinates of the clustering results are calculated statistically.Then segmented equalization is carried out according to the statistical results in the original infrared thermal image,and the segmented equalization images are coded by rainbow encoding method.Experimental results show that the infrared thermal image encoded by this method has rich color and detail,and has a better sense of hierarchy,which is convenient for human eye observation and analysis.