Metabolic heat, the product following the metabolism of cells, is closely related to the pathological information of living organisms, which means there are strong connections between the heat distribution and the pathological state of the living organism. The mathematical function δ is introduced in the classical Pennes bioheat transfer equation as a point heat source, and by simplifying the boundary condition, a bioheat transfer model is established. Based on the temperature distribution of the human body surface, the q−r curve of heat intensity q varying with depth r is acquired while combining the ﬁtting method of the Lorentz curve. According to 34,977 clinical confirmed cases and the corresponding classified statistics, diagnostic criteria (for breast diseases) for judging diseases by the q−r curve are proposed. The P-value of our statistics is <0.05, which means our classified statistics are reliable. Six typical clinical examinations are performed, and the diagnosis results are very consistent with those of B-ultrasonic images, molybdenum target x-ray, and pathological examination, which suggests that the method of diagnosing diseases with a q−r curve has very good prospects for application. It is radiation free and noninvasive to the human body.