Background <p>Previous studies have suggested an association between tinnitus and various psychiatric phenotypes; however, the causal relationship and underlying pathways remain unclear.</p> Methods <p>We used univariable Mendelian randomization (MR) to investigate the bidirectional causality between tinnitus and 8 psychiatric phenotypes. Multivariable MR was employed to control for potential confounders. Extensive sensitivity analyses were conducted to validate our results. Additionally, two-step MR and genetic pleiotropy analyses were performed to explore potential pathways.</p> Results <p>Genetic susceptibility to tinnitus was significantly associated with an increased risk of major depressive disorder (MDD) (OR = 1.07, 95% CI = 1.03–1.11). Multivariable MR demonstrated that the impact of tinnitus on MDD was independent of potential confounders. No causal effects were observed from the eight psychiatric phenotypes on tinnitus; likewise, no causal effects were found from tinnitus on the remaining seven psychiatric phenotypes. Two-step MR analysis revealed that hearing difficulties with background noise and insomnia mediated the effect of tinnitus on MDD, with mediation proportions of 20.0% (0.0%–40.1%) and 31.4% (0.9%–61.9%), respectively. Genetic pleiotropy analysis identified a global genetic correlation between tinnitus and MDD (r_g = 0.36–0.48). PLACO analysis revealed 66 shared loci, primarily enriched in the cerebellum and cerebellar hemisphere. FUMA identified two independent genomic risk loci (rs58825580 and rs729437) and mapped 91 pleiotropic genes, which were mainly enriched in nucleosome and chromatin-remodeling processes, and immune response dysregulation.</p> Conclusions <p>This study provides genetic evidence that genetic susceptibility to tinnitus has a causal effect on the risk of MDD, partially mediated by hearing difficulties and insomnia. Our genetic pleiotropy analysis further highlights the shared genetic architecture between tinnitus and MDD.</p>