We conducted a review of systematic reviews (SRs) and randomized-controlled trials (RCTs) to analyze efficacy and safety of cannabis-based medication in patients with mental disorders. Five data bases were systematically searched (2006—August 2018); 4 SRs (of 11 RCTs) and 14 RCTs (1629 participants) were included. Diagnoses were: dementia, cannabis and opioid dependence, psychoses/schizophrenia, general social anxiety, posttraumatic stress disorder, anorexia nervosa, attention-deficit hyperactivity disorder, and Tourette`s disorder. Outcome variables were too heterogeneous to conduct a meta-analysis. A narrative synthesis method was applied. The study quality was assessed using the risk-of-bias tool and SIGN-checklists. THC- and CBD-based medicines, given as adjunct to pharmaco- and psychotherapy, were associated with improvements of several symptoms of mental disorders, but not with remission. Side effects occurred, but severe adverse effects were mentioned in single cases only. In order to provide reliable treatment recommendations, more and larger RCTs with follow-up assessments, consistent outcome measures and active comparisons are needed.