Abstract When there exist diseases or functional changes in a certain part of the human body, the speed of blood flow and cell metabolism will change correspondingly, which will lead to the thermal variation in this area. To find out the relation between disease and heat distribution, a suitable bio-heat transfer model is established in this paper. Based on the infrared thermal image of human body surface, the q–r characteristic curve of heat intensity varying with depth is acquired combining the fitting method of Lorentz curve. According to a large number of clinical cases and statistics, the diagnostic criteria judging diseases by q–r characteristic curve are proposed. Several clinical practices are performed and the diagnosis results are very consistent with those of molybdenum target (MT) X-ray and B-ultrasonic images. It is a radiation-free green method with rapid diagnostic procedure and accurate diagnosis result.