The aim of this study was to appraise the effects of motor imagery on the functional performance improvement among total knee arthroplasty patients systematically. We hypothesized a relatively greater recovery in the motor imagery group.Medline (Ovid), Embase and Cochrane Controlled Register of Trials (CENTRAL) were searched from inception to October 1st, 2021. We included randomized controlled trials evaluating the effects of motor imagery on the functional recovery among total knee arthroplasty patients. Measurements included range of motion, strength intensity, Visual Analogue Scale, Time Up and Go Test, Oxford Knee Score, Western Ontario and McMaster Universities Osteoarthritis Index, all of which were evaluated before and after intervention. Mean differences (MD) or standard mean differences (SMD) and 95% confidence intervals (CI) were calculated. The Cochrane risk of bias tool was used to assess the risk of bias.Six studies with 168 patients were included for the meta-analysis. The SMD of strength intensity was increased (SMD = 0.90, 95% CI = [0.47]-[1.32], P < 0.001). The SMD of Visual Analogue Scale was reduced (SMD =  - 0.91; 95% CI = [- 1.29]-[- 0.52], P < 0.001). The SMD of Time Up and Go Test was reduced (SMD =  - 0.56, 95% CI = [- 0.94]-[- 0.19], P = 0.003). The MD of Oxford Knee Score was slightly increased (MD = 0.79-point, 95% CI = [- 0.31]-[1.88], P = 0.159). The outcomes of range of motion, Western Ontario and McMaster Universities Osteoarthritis Index were described according to the original data.Compared with control therapy, motor imagery in the intervention group achieved an effective treatment for strength enhancement, pain reduction and physical activities improvement. More large-scale, prospective researches are needed in the future.The PROSPERO trial registration number is CRD42021250996.© 2022. The Author(s).