Virtual reality-based rehabilitation (VRBR) is being used in rehabilitation after total knee arthroplasty (TKA). This meta-analysis assesses the effect of immediate and/or maintained effect of VRBR (alone or in combination with conventional therapy (CT) in TKA recovery. To perform this meta-analysis, we searched in PubMed, SCOPUS, WOS, CINAHL, and PEDro until 5 February 2023. We included randomized controlled trials that assessed the effect of VRBR in improving knee pain and function, dynamic balance, range of motion (ROM), and strength, among others, after TKA. The pooled effect was calculated with Cohen’s standardized mean difference (SMD) with a 95% confidence interval (95% CI). Twelve studies, providing data from 997 participants (69.9 ± 8.1 years old), were included. At post-immediate assessment, VRBR is effective in improving knee pain (SMD = −0.36; 95% CI −0.56 to −0.17), knee function (SMD = −0.51; 95% CI −0.75 to −0.27), dynamic balance (SMD = −0.59; 95% CI −1.02 to −0.15), knee flexion ROM (SMD = 0.4; 95% CI 0.17 to 0.62), and extension strength (SMD = 0.43; 95% CI 0.19 to 0.68). Our findings showed that immersive and non-immersive VRBR is effective to be used after TKA, and the effect on knee pain and function may be maintained over 3 or 6 months. A high level of satisfaction and adherence to VRBR was reported.