Objective To evaluate the efficacy and safety of the midvastus approach in total knee arthroplasty. Methods We searched The Cochrane Library, EMBASE, PubMed, and CBM to identify clinical controlled trials comparing the midvastus approach with the medial parapatellar approach in total knee arthroplasty. The quality of the included studies was critically assessed and the data analyses were performed by the Cochrane Collaboration’s RevMan 5.0. Results Eight studies were included, involving 4 randomized controlled trials (RCTs), 2 quasi-RCTs and 2 non-RCTs. Meta-analyses showed that rate of lateral retinacular release (RR=0.75, 95%CI 0.52 to 1.08, P=0.12), range of motion at 6 week postoperation (MD=2.65, 95%CI –1.20 to 6.50, P=0.18), operation time (MD=1.04, 95%CI –3.50 to 5.58), and adverse events postoperation (OR=1.04, 95%CI 0.43 to 2.52, P=0.94) were similar between the midvastus approach and the medial parapatellar approach in total knee arthroplasty. One study showed that there was no significant difference in blood loss interoperation between the two approaches, and two showed that the midvastus approach had less blood loss interoperation than the medial parapatellar approach. There was no statistical analysis about the blood loss interoperation in one study. For the time of straight leg raise, there was no significant difference in two studies. But in one study, it showed that patients needed longer time for straight leg raise. Conclusion Based on the current evidence, the midvastus approach for total knee arthroplasty is as safe and effective as the medial parapatellar approach, but blood loss interoperation and time of straight leg raise are not decided. Due to the poor quality of the included trials, more high-quality RCTs are needed.