There is no comprehensive meta-analysis of randomised trials examining the effects of Eye Movement Desensitization and Reprocessing (EMDR) on post-traumatic stress disorder (PTSD) and no systematic review at all of the effects of EMDR on other mental health problems. We conducted a systematic review and meta-analysis of 76 trials. Most trials examined the effects on PTSD (62%). The effect size of EMDR compared to control conditions was g = 0.93 (95% CI: 0.67-.18), with high heterogeneity (= 72%). Only four of 27 studies had low risk of bias, and there were indications for publication bias. EMDR was more effective than other therapies (g = 0.36; 95% CI: 0.14-0.57), but not in studies with low risk of bias. Significant results were also found for EMDR in phobias and test anxiety, but the number of studies was small and risk of bias was high. EMDR was examined in several other mental health problems, but for none of these problems, sufficient studies were available to pool outcomes. EMDR may be effective in the treatment of PTSD in the short term, but the quality of studies is too low to draw definite conclusions. There is not enough evidence to advise it for the use in other mental health problems.