Data from clinical trials of couple therapy for depression have never been subjected to systematic analyses. We performed a meta-analysis of eight controlled trials involving 567 subjects. No difference was found on depressive symptoms between couple therapy and individual psychotherapy. Relationship distress was significantly reduced in the couple therapy group. Too few data are available for comparisons with drug therapy and no treatment. The findings are weakened by small sample sizes, assessments at the end of treatment or short follow-up, unclear sample representativeness, heterogeneity among studies. The mediating role of improvement in quality of couple relationships is not supported by data. However, it has not been adequately tested. Evidence on efficacy of couple therapy as a treatment for depression is inconclusive. The evidence for improvement in couple relationships may favor the choice of couple therapy when relational distress is a major problem.