The conventional athlete injury motion capture method mainly uses Orthoload optical capture standard calibration database to calculate the load on the injured joint, which is easily affected by the change of the sagittal plane coefficient of the contact moment, resulting in a high deviation of the knee flexion moment (KFM) of the injured joint. Therefore, a multisource sensing signal-based athlete injury motion capture method is proposed. By utilizing multisource sensing signals for adaptive detection of human posture domain, an athlete injury motion capture classifier was designed to achieve athlete injury motion capture. The experimental results show that the designed multisource sensing signal capture method for athlete injury movements captures the joint bending moment KFM of the injured joint and fits it with the actual joint bending moment KFM of the injured joint, which is reliable and has certain application value. It has made a certain contribution to reducing athlete sports risks and formulating reasonable repair plans.