Transfemoral amputees use knee-ankle prostheses to restore impaired locomotion. Active prostheses can potentially overcome the limitations of passive ones by better supporting missing muscle function with additional motor power. The control algorithms of such embedded motors should be tailored to understand the users locomotive intention and translate them into the required locomotion similar to that of an able-bodied individual. For individuals with transfemoral amputation, the control algorithm should produce the desired locomotion by controlling an active knee and ankle joint to generate appropriate knee angle, knee moment, ankle angle and ankle moment. Machine learning models could be utilized to develop control algorithms for active prosthesis. In this paper, a random forest strategy is proposed for the continuous prediction of the angles and moments of the knee and ankle during walking. Experimentally obtained thigh motion data were provided as the input to the random forest model. It is shown that, for level ground walking at self-selected speed, the proposed method could predict the angles and moments of the knee and ankle with high accuracy (mean R ² value of 0.995 for knee angle, 0.951 for knee moment, 0.966 for ankle angle, and 0.941 for ankle moment). The proposed strategy shows potential for continuously controlling an active ankle-knee prosthesis for a transfemoral amputee, whose thigh angular motion can be used to infer the required prosthetic moments and angles.