Background: Nowadays, antidepressants still are the mainstay of treatment for depression in Parkinson’s disease (PD) but some recent studies report that medication might aggravate motor symptoms in PD patients. This meta-analysis aims to assess the effect of non-pharmacological treatments for depression in patients with PD. Materials and Methods: Only randomized controlled trials (RCTs) were included. The participants were PD patients with comorbid depression (dPD). The interventions had the equivalent effect of non-pharmacological treatments alone compared with control(s). Scores of depression scale were selected as the primary outcome, while scores of Unified Parkinson’s Disease Rating Scale part III and the incidence of side effects were the secondary outcome. The statistics were pooled and presented as weighted mean differences (WMDs), standardized mean differences (SMDs), or risk ratios (RRs) with their 95% confidence intervals (CIs). Results: Fifteen articles were eventually included; twelve studies reported on repetitive transcranial magnetic stimulation (rTMS) and three used cognitive behavioral therapy (CBT). Other interventions failed to have qualified studies. Our data indicated that both rTMS and CBT could significantly improve depression scores in a short term (SMD ¼0.621, 95% CI [0.964, 0.278]; SMD ¼1.148, 95% CI [1.498, 0.798], respectively). In addition, rTMS could alleviate motor symptom (WMD ¼2.617, 95% CI [4.183, 1.051]) and was relatively safe (RR ¼ 1.054, 95% CI [0.698, 1.592]). Conclusion: Our data suggest that rTMS can safely alleviate depression and motor symptoms in dPD at least for a short period. Moreover, compared with clinical monitoring, CBT can improve depressive symptoms.