In order to monitor and analyze the state of the body during movement more effectively, an infrared thermal radiation image acquisition system based on computer simulation was developed. A simulation model containing the thermodynamic characteristics of the human body was constructed, which could simulate the change of body temperature under different movement intensity and environmental conditions. A high precision infrared thermal imager is used to collect infrared thermal radiation images in simulated motion state. The infrared thermal radiation images are preprocessed by image processing techniques such as filtering, enhancement and edge detection. Then, the image features are extracted by machine learning algorithm, and the correlation model between body temperature distribution and motion state is established. In the actual sports fitness scene, the developed system is used to monitor the athletes in real time, and compared with the traditional monitoring means to verify the accuracy and practicability of the system. The experimental results show that the developed system can accurately reflect the body temperature changes under different exercise intensities, and has higher real-time and accuracy than the traditional monitoring means. The system can also predict the fatigue state and the potential risk of sports injury, which provides important reference information for sports fitness.