OBJECTIVE: To evaluate the effectiveness of VR therapy (VRT) for symptoms related to social anxiety disorder (SAD), namely fear and avoidance of social interactions and performance situations (FASIP), fear of negative evaluation (FNE), anxiety and depression, a systematic review and meta-analysis were performed. METHODS: Medline, PubMed, Science Direct, Web of Science, CINAHL, PsychINFO and Scopus were searched to include randomised controlled trials of VRT for SAD that met the criteria. A total of 15 RCTs with 720 participants published between 1998 and 2022 were included. Hedge's g with a 95 % confidence interval (CI) was adopted to compute the effect sizes. RESULTS: Results showed no difference between the effect of VRT and CBT on FASIP, FNE, anxiety and depression and a large effect size for VRT versus the waitlist control group on FASIP (g = -1.170, 95 % CI: -2.056-0.283; p < 0.010). The moderator analysis demonstrated that VRT was superior to the controlled group in addressing FASIP, FNE and anxiety when the sample size was smaller than 50 and the number of sessions was five or fewer. LIMITATIONS: Differences in hardware, software and intervention duration for VRT across studies. CONCLUSION: This study confirmed the feasibility of VRT in alleviating the FASIP in patients with SAD, with the waitlist control group as a comparison. However, the effectiveness of VRT was not significant in FASIP, FNE, anxiety and depression compared to cognitive behavioural therapy (CBT). Additional social interaction scenarios should be developed in VRT, standardised hardware should be used and the proper length of exposure time to VR should be determined to enhance the efficacy of VRT.