Conventional instrumentation systems have limited accuracy in determining the crucial landmarks needed for alignment in total knee arthroplasty (TKA). Given this, the image-based navigation system was introduced to improve the accuracy of implantation of components into the femur and tibia. PubMed, EMBASE, Web of Science, and Evidence-Based Medicine databases were electronically searched to identify eligible studies published until October 2008. A systematic review and meta-analysis of 6 randomized/quasi-randomized controlled trials that compared image-based navigation and conventional techniques was conducted. The operative time was longer in the navigation group in 3 studies. Moreover, there was a higher rate of achieving mechanical leg axis within the range of 3° deviation in patients undergoing navigated TKA. However, all studies between the 2 groups were similar in range of motion, knee scores, and postoperative complication rates at the last follow-up. Overall, these short-term follow-up trials show that there were similar early clinical outcomes between image-based navigation and conventional techniques.