The purposes of this study were (1) to compare lower limb alignment measurements between radiographs and computer-assisted surgery and (2) to evaluate the discrepancy in lower limb alignment between computer-assisted surgery with a high tibial osteotomy protocol and computer-assisted surgery with a total knee arthroplasty (TKA) protocol in the same knee. Seventy-one TKAs were performed on patients with primary osteoarthritis using computer-assisted surgery. Preoperative lower limb alignment was measured using the mechanical axis during bipedal, weight-bearing, whole-leg anteroposterior radiography (measure 1). The intraoperative mechanical axis was measured with computer-assisted surgery according to the high tibial osteotomy protocol before joint exposure (measure 2). After changing the software and joint exposure, the intraoperative mechanical axis was measured with computer-assisted surgery according to TKA protocol (measure 3). After final TKA implantation, the lower limb mechanical axis was measured with computer-assisted surgery following the TKA protocol (measure 4). Postoperative lower limb alignment was measured using the mechanical axis on whole-leg standing anteroposterior radiographs (measure 5). The mechanical axis and median value from each group were compared. Factors affecting the mechanical axis measurement were also analyzed. The difference in the mechanical axis between measures 1 and 2, measures 1 and 3, and measures 2 and 3 was significant (P<.0001, <.0001, and =.0007, respectively). The difference between measures 4 and 5 was also significant (P<.0001). Factors affecting the mechanical axis measurement, such as age, height, weight, and range of motion, showed no correlation (R(2)=.07244 and adjusted R(2)=.01622). The pre- and postoperative radiological measurements of limb alignment using the mechanical axis were different from the intraoperative measurements with navigation.