This study aimed to assess the efficacy of various music therapy interventions in ameliorating depressive symptoms in dementia patients, utilizing a network meta-analysis approach. We rigorously selected randomized controlled trials focused on music therapy for dementia with depressive symptoms from major electronic databases. The primary outcome measured was the impact on depressive symptoms, with the secondary outcome evaluating dropout rates across different intervention groups and standard care control groups. The research protocol has been duly registered with PROSPERO (Registration ID: CRD42023393059). Our network meta-analysis incorporated 14 randomized controlled trials involving a total of 1080 participants and examined a range of interventions, including active music therapy, listening to music, rhythmic music therapy, singing, and tailored music interventions. The analysis revealed that active music therapy combined with singing emerged as the most effective intervention, demonstrating a significant improvement in depressive symptoms in dementia patients (Standardized Mean Difference [SMD] = −0.89, 95% Confidence Interval [CI]: −1.48 to −0.30). In contrast, listening to music alone showed a smaller effect (SMD = −0.26, 95% CI: −0.71 to 0.20). This study was particularly noteworthy for not showing higher dropout rates compared to standard care, indicating its feasibility and acceptability in clinical settings. The findings of our study indicate that active music therapy combined with singing is an effective approach to reducing depressive symptoms in dementia patients, potentially due to enhanced social interaction. These results offer new perspectives for dementia care, suggesting a promising direction for further research and clinical application.