Femoral condyle motion of the knee is generally reported using a morphological trans-epicondyle axis (TEA) or geometric center axis (GCA) in the investigation of the knee kinematics. Axial rotation of the femur is recognized as a characteristic motion of the knee during flexion, but is controversial in the literature. This study investigated the biomechanical factors that could be associated to the axial rotations of the femur using both physiological and morphological measurement methods. Twenty healthy knees were investigated during a weightbearing flexion from 0° to 120° at a 15° increment using an imaging technique. A 3D model was constructed for each knee using MR images. Tibiofemoral cartilage contact points were determined at each flexion position to represent physiological knee motion. The contact distance on each condyle was measured between consecutive contact points. The TEA and GCA were used to measure morphological anteroposterior translations of the femoral condyles. The differences between the medial and lateral condyle motions were used to calculate the physiological and morphological axial rotations of the femur. Both the physiological and morphological methods measured external rotations of the femur at low flexion range (0°-45°) and minimal rotations at higher flexion angles. However, the morphological method measured larger posterior translations of the lateral femoral condyle than the medial condyle (p < 0.05), implying a medial pivoting rotation; in contrast, the physiological method measured larger contact distances on the medial condyle than on the lateral condyle (p < 0.05), implying a lateral pivoting rotation. These data could provide useful references for future investigation of kinematics of the knee before and after surgical repair, such as using total knee arthroplasty.