Objective To compare the temperatures at different parts of human body measured by two thermometers, and to provide evidences for clinical temperature measurement sites and tools. Methods A self-controlled descriptive study was conducted. A non-contact infrared thermometer was used to measure the temperature of the forehead, temporal artery, middle of the neck, lateral of the neck, medial corner of the eye, posterior earlobe, and the ear canal, with a distance of 1.5 cm and 5.0 cm respectively. Meanwhile, the axillary temperature was measured with a mercury thermometer and served as a reference. Results A total of 76 patients were included. The overall difference in body temperatures measured at 7 sites by the non-contact infrared thermometer was significant, and the measurements had significant difference between the 1.5 cm and the 5 cm subgroups (P＜0.01 for all). The average temperature of the ear canal was the highest, while the temperature of forehead was the lowest. For patients without a fever, the temperature measured at the forehead, and temporal artery had significant difference with axillary temperature（P<0.05,P<0.01). For patients with a fever, the temperatures taken at the forehead, temporal artery and the neck were significantly different from axillary temperature（P<0.05,P<0.01), with the disparity between forehead and axillary temperature as large as 0.62℃ (P<0.01). Ear canal temperature was the closest to axillary temperature (P＞0.05). Conclusion The temperature measured at the ear canal is closer to that taken at the axilla. It is recommended to use the ear canal as the measurement site for fever screening. Keeping a distance of 5.0 cm between the thermometer and skin could reduce the risk of contact while ensuring accuracy. However, in the case of ear canal measurement, the distance can be appropriately shortened to better take aim.