Soft tissue balancing remains the most subjective and most artistic of current techniques in total knee arthroplasty. The flexion gap is traditionally measured at approximately 45 degree of hip flexion and 90 degree of knee flexion on the operation table. Despite of aiming equal joint gaps or tensions in flexion and extension, influence of the thigh weight on the flexion gap has not been documented. Therefore, the purpose of this study was to examine the flexion gaps in the 90–90 degree flexed position and the traditional 45–90 degree flexed position of hip-knee joints. Thirty patients with osteoarthritic knee underwent total knee arthroplasty. After the PCL sacrifice, soft tissue releases, and bone cuts. Biomechanical properties of the soft tissue were obtained during the surgery, using the specially designed system. The system consists of two electric load cells in the tensioning device, digital output indicators, and an XY plotter. Load displacement curves were obtained in extension and in flexion. 160N was applied to open the joint gaps in the traditional 45–90 degree flexed position and the 90–90 degree flexed position of hip-knee joints. The flexion gap in the 90–90 degree flexed position of hip-knee joints was 2.1±1.2mm wider than that in the traditional 45–90 degree flexed position of hip-knee joints. The flexion gap had significant difference between the two different hip flexion angles. To avoid the influence of the thigh weight and obtain equal joint gaps or tensions in flexion and extension, the flexion gap should be checked in the 90–90 degree flexed position of hip-knee joints. Interestingly, the stiffness of curves obtained from the lateral in flexion is 1/3 lower than the other three. Therefore, it is very difficult to match these four. The effect of patellar position on soft tissue balancing in TKA is also under debate. We developed the digital tensor system to measure the load (N) and the distance (mm) of extension and flexion gaps in medial and lateral compartment separately with setting of femoral component trial. The gap load and distance in extension and flexion position of PS and CR TKA in both patella everted and reset position were measured. Thirty-four patients who underwent primary TKA for medial type osteoarthritis using medial parapatellar approach were included. The load was measured at the gap distance, which is equal to the sum of implants including polyethylene insert. In extension, there was no significant difference between the load in patella everted and reset position in both PS-TKA and CR.-TKA. In flexion, there was a significant decrease of the load, which is comparable to the increase of gap distance of approximately 2mm, by resetting the patella from eversion in PS-TKA. There was, however, no significant difference in CR-TKA by resetting the patella. There was no significant difference in the ratio of medial / lateral load in both PS-TKA and CR.-TKA. Soft tissue balancing of PS-TKA with medial parapatellar approach should be performed after resetting the patella. It is still unclear whether we can adjust these materials precisely and constantly or not