Recent evidence suggests that vegetarian and vegan diets may increase the risk and symptoms of depression, a mental health condition affecting 350 million people globally. We aimed to systematically review the literature on the associations between vegetarian and/or vegan diets and the risk or symptoms of depression using evidence from both observational and intervention studies. We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, with pre-specification of all methods. A systematic search for relevant papers was performed on Medline and Embase, Web of Science and the Cochrane Library for cohort, case-control, cross-sectional studies or randomised controlled trials examining associations between a vegetarian or vegan diet and depression in adults. Three independent reviewers extracted data and assessed risk of bias using the National Heart, Lung, and Blood Institute of the National Institutes of Health for Quality Assessment of Observational Cohort and Cross-Sectional Studies and Controlled Studies. Evidence was tabulated according to the type of diet analysed as vegetarian, vegan or both and narratively synthesised. A total of 23 studies (18 cross-sectional, three prospective cohort and two randomised controlled trials) with 25 study outcomes were eligible for inclusion in this review. Conflicting evidence was found on the association between vegetarian or vegan diets and depression. Eleven (44%) of the outcomes indicated that vegetarian and vegan diets were associated with higher rates of depression, while seven (28%) outcomes revealed beneficial effects of the diets on depression. Seven (28%) outcomes found no association between vegetarian and vegan diets and depression, although two of these studies found a higher risk of depression in some groups. The quality of evidence was rated as good for four of the studies with the remaining 19 studies rated as fair. The evidence on the effect of vegetarian and vegan diets on depression is contradictory, possibly due to the heterogeneity of the studies analysed. Further research, including longitudinal and intervention studies, is required to resolve this observation.