Background We aimed to synthesize the available evidence on the relative efficacy and acceptability of specific treatments for persistent depressive disorder. Methods We searched several databases up to January 2013 and included randomized controlled trials that compared acute pharmacological, psychotherapeutic, and combined interventions with each other or placebo. The outcome measures were the proportion of patients who responded to (efficacy) or dropped out from (acceptability) the allocated treatment. Data synthesis was performed with network meta‐analysis. Results A network of 45 trials that tested 28 drugs included data from 5,806 and 5,348 patients concerning efficacy and acceptability, respectively. A second network of 15 trials that tested five psychotherapeutic and five combined interventions included data from 2,657 and 2,719 patients concerning efficacy and acceptability, respectively. Among sufficiently tested treatments, fluoxetine (odds ratio (OR) 2.94), paroxetine (3.79), sertraline (4.47), moclobemide (6.98), imipramine (4.53), ritanserin (2.35), amisulpride (5.63), and acetyl‐l‐carnitine (5.67) were significantly more effective than placebo. Pairwise comparisons showed advantages of moclobemide (2.38) and amisulpride (1.92) over fluoxetine. Sertraline (0.57) and amisulpride (0.53) showed a lower dropout rate than imipramine. Interpersonal psychotherapy with medication outperformed medication alone in chronic major depression but not in dysthymia. Evidence on cognitive behavioral analysis system of psychotherapy plus medication was partly inconclusive. Interpersonal psychotherapy was less effective than medication (0.48) and cognitive behavioral analysis system of psychotherapy (0.45). Several other treatments were tested in single studies. Conclusions Several evidence‐based acute pharmacological, psychotherapeutic, and combined treatments for persistent depressive disorder are available with significant differences between them.