The purpose of this work is to describe a computational framework for predicting total knee arthroplasty loads which are necessary for accurate preclinical testing of implant designs. Inputs required include patient knee joint kinematics, and implant type, size, and physiological alignment. Computational models used in the framework include the calculation of knee joint kinematics and kinetics, prediction of the contact path, a model to determine muscle forces, and a force model to obtain parametric solutions for implant forces. The resulting knee implant forces have been validated in two studies, and in both the model accurately predicted differences in knee joint loading. To date, implant contact forces have been predicted for 35 patients with four different implant types. Forces have been calculated for walking, chair, and stair activities.