The COVID-19 pandemic has spread all over the world. People go to public or crowded areas (i.e., schools, universities, hospitals, and government agencies), they take a lot of time to be checked the fever symptoms because of coronavirus. Therefore, this paper presents a method to automatically detect the body temperature by distance based on the recursive least square estimation. An infrared thermal camera is utilized to measure both human and environmental temperatures in real-time within a two-meter distance. The recursive least square approach is applied to estimate parameters for these correct temperatures. A microcontroller is integrated to read, compute, and send the measured temperatures to both web browsers and smartphones using the message queuing telemetry protocol. Moreover, the module of radio frequency identification is utilized for identification of the personal information. To validate our proposed temperature measurement system, fifteen male healthy students are invited to record their body temperature. The experimental result showed that our proposed approach was the correct temperature compared with the commercial device ( 37±0.17∘C ). However, our proposed system is more stable than the commercial device: the standard deviation of the commercial device and ours is 0.41 C and 0.09°C, respectively. The measured temperature of each person is monitored and stored in the cloud. It is easily accessed by web browsers and smartphones. In addition, our proposed system can show a warning if the measured temperature is greater than the threshold. This work promises to automatically initial selection for suspected cases of COVID-19 disease to reduce the infection of this pandemic.