PurposeTo compare the clinical efficacy of single‐bundle versus double‐bundle reconstruction of the medial patellofemoral ligament (MPFL) for recurrent patellar dislocation (RPD) regarding knee function scores, postoperative complications, and imaging assessments.MethodsA computerized search of PubMed, Cochrane Library, Embase, China Biomedical Literature Database (CBM), China National Knowledge Network (CNKI), and VIP Database was performed for single‐bundle versus double‐bundle reconstruction of the medial patellofemoral ligament for treatment of RPD. Randomized controlled trials (RCTs) were evaluated for quality using the risk‐of‐bias evaluation tool recommended by the Cochrane Collaboration Network, and Cohort studies (CSs) were assessed using the Newcastle‐Ottawa Scale (NOS) scale. Meta‐analysis was performed using RevMan 5.3 software and STATA 16.0.ResultsThirteen studies were included, four randomized controlled studies, and nine cohort studies. The level of evidence for the four randomized controlled studies was Ⅰ, and the nine cohort studies were Ⅲ. A total of 862 (891 knees) patients were included, of which 448 (465 knees) underwent double‐bundle MPFL reconstruction and 414 (426 knees) underwent single‐bundle MPFL reconstruction. Kujala score (MD = 2.06, 95% confidence interval [CI] [0.11, 4.01], p < 0.05), Tegner score (MD = 0.39, 95% CI [0.11, 0.68], p < 0.05), International Knee Documentation Committee (IKDC) score (MD = 4.88, 95% CI [1.46, 8.31], p < 0.05), and postoperative recurrence instability (odds ratio [OR] = 0.12, 95% CI [0.04, 0.44], p < 0.05) were better in the double‐bundle group than in the single‐bundle group. Lysholm score (MD = 0.86, 95% CI [−0.76, 2.48], p = n.s), patellar tilt angle (MD = −0.22, 95% CI [−0.54, 0.10], p = n.s), patellar lateral shift rate (MD = −0.16, 95% CI [−0.41, 0.09], p = n.s), congruence angle (MD = 0.06, 95% CI [−0.41, 0.52], p = n.s), postoperative knee pain (OR = 0.39, 95% CI [0.14, 1.11], p = n.s), and additional postoperative surgical treatment (OR = 0.20, 95% CI [0.01−6.25], p = n.s) had no statistically significant differences.ConclusionsDouble‐bundle reconstruction of the medial patellofemoral ligament for RPD was superior to single‐bundle reconstruction in both knee function scores and postoperative recurrent patellar instability, and double‐bundle reconstruction of the medial patellofemoral ligament for RPD had better clinical outcomes.Level of EvidenceLevel Ⅲ, Ⅰ and Ⅲ studies.