Yoga has become a popular approach to improve emotional health. The aim of this review was to systematically assess and meta‐analyze the effectiveness and safety of yoga for anxiety. Medline/PubMed, Scopus, the Cochrane Library, PsycINFO, and IndMED were searched through October 2016 for randomized controlled trials (RCTs) of yoga for individuals with anxiety disorders or elevated levels of anxiety. The primary outcomes were anxiety and remission rates, and secondary outcomes were depression, quality of life, and safety. Risk of bias was assessed using the Cochrane tool. Eight RCTs with 319 participants (mean age: 30.0–38.5 years) were included. Risk of selection bias was unclear for most RCTs. Meta‐analyses revealed evidence for small short‐term effects of yoga on anxiety compared to no treatment (standardized mean difference [SMD] = −0.43; 95% confidence interval [CI] = −0.74, −0.11; P = .008), and large effects compared to active comparators (SMD = −0.86; 95% CI = −1.56, −0.15; P = .02). Small effects on depression were found compared to no treatment (SMD = −0.35; 95% CI = −0.66, −0.04; P = .03). Effects were robust against potential methodological bias. No effects were found for patients with anxiety disorders diagnosed by Diagnostic and Statistical Manual criteria, only for patients diagnosed by other methods, and for individuals with elevated levels of anxiety without a formal diagnosis. Only three RCTs reported safety‐related data but these indicated that yoga was not associated with increased injuries. In conclusion, yoga might be an effective and safe intervention for individuals with elevated levels of anxiety. There was inconclusive evidence for effects of yoga in anxiety disorders. More high‐quality studies are needed and are warranted given these preliminary findings and plausible mechanisms of action.