The aim of this study is to assess the radiological outcome of conventional techniques versus computer-navigated surgery for total knee arthroplasty. Ninety patients with knee arthritis were prospectively randomized into 3 groups: conventional technique: extramedullary (EM) and intramedullary (IM) tibia guide versus computer navigation surgery (CAS). Two surgeons performed all procedures. Standardized long leg coronal and sagittal x-rays were evaluated by a blinded assessor. Our results showed that CAS had greater consistency and accuracy in implant placement. In the coronal view, 93.3% in the CAS group had better outcomes compared with EM (73.4%) and IM (60.0%). In the sagittal axis, 90.0% CAS also had better outcomes compared with EM (63.3%) and IM (76.7%). Computer-navigated total knee arthroplasty helps increase accuracy and reduce "outliers" for implant placement.