The infrared thermal image distribution and mathematical mean of human′s breast region were observed. Their value to the diagnosis of breast diseases is analyzed. 125 patients who met the conclusion criteria in the outpatient department of the treating potential disease center of Liuzhou Traditional Chinese Medicine Hospital from June 2016 to January 2017 were selected as research objects. The mathematical mean of infrared thermal imaging in the projection region of double mammary gland was observed by a digital medical infrared thermal imager. At the same time, breast ultrasound and mammography were performed on all objects. The mathematical mean of infrared thermal imaging of each object, their distribution rule and the results under different examination items were recorded. The correlation and difference of the results were analyzed statistically. The results showed that there was a positive correlation between the mathematical mean of infrared thermal imaging and the degree of breast lesions. The results of breast ultrasound and mammography were compared with the corresponding mathematical difference of thermal imaging. Their difference was not significant statistically (P>0.05). Therefore, there was a correlation between the mathematical mean of infrared thermal imaging of breast region and different breast diseases. Among them, the mathematical mean of infrared thermal imaging of breast cancer was highest and that of normal breast was lowest. To some extent, the mathematical mean of infrared thermal imaging was more sensitive to the breast diseases than the breast ultrasound in diagnosis.