Psychotherapies for social anxiety disorder (SAD) encompass various protocols and delivery formats; consequently, an investigation into the relative efficacy of different treatments is needed. This study aimed to compare the efficacy of different common psychotherapies for SAD and identify the most effective psychotherapies for SAD. Using a Bayesian network meta-analysis, we compared the relative efficacy of six common psychotherapies: cognitive behavioral therapy (CBT), cognitive restructuring (CR), exposure therapy (ET), psychodynamic therapy (PDT), interpersonal therapy (IPT), and mindfulness-based interventions (MBI). Sensitivity analyses were conducted to test the robustness of the results. This network meta-analysis included 92 studies up to June 2024, covering 90 randomized controlled trials. The findings suggested that, compared with the waitlist control group, all included psychotherapies showed varying degrees of efficacy in reducing SAD symptoms severity. Specifically, CBT was generally the most effective psychotherapy. The Clark and Wells CBT protocol showed the highest efficacy (SMD: 1.42, 95 % CrI: [1.14, 1.70]), while Andersson and Carlbring's Internet-based CBT ranked highly (SMD: 1.15, 95 % CrI: [0.87, 1.42]). PDT was the most effective non-CBT treatment (SMD: 0.98, 95 % CrI: [0.61, 1.36]). Guided therapies appeared more effective than unguided ones, and Internet-based treatments were comparable to face-to-face therapies. This study can aid clinicians in optimizing treatment choices and selecting appropriate formats when treating patients with SAD.