Background This study examined the prospective association of workplace social capital (WSC) with major depressive episode (MDE) among Japanese employees. Methods A 3-year prospective cohort study was conducted among 1058 employees from a private thinktank company who participated in a baseline survey; after excluding those with MDE in the past 12 months, 929 were followed up. WSC at baseline was measured using a 3-item scale. MDE was assessed at baseline and at follow-up every year, by using a web-based, selfadministered version of the WHO Composite International Diagnostic Interview (WHO-CIDI) 3.0 depression section, based on the Diagnostic and Statistical Manual of Mental Disorders Fourth Edition: Text Revision (DSM-IV-TR)/DSM-5 criteria. Cox discretetime hazards analyses were used to estimate HRs and 95% CIs, adjusting for covariates. Results A group with middle-level WSC scores had the lowest risk of MDE after being fully adjusted (HR 0.34, 95% CI 0.14 to 0.84, p=0.02). The relationship between WSC and MDE was U-shaped, although a nonlinear model fit better than a linear model, with only marginally statistical significance (p=0.06). Dichotomised and continuous variables of WSC scores were significantly and negatively associated with MDE (p=0.03 and p<0.01, respectively). Conclusions The current study replicated a previous finding from Finland that WSC was a protective factor of the onset of MDE in Japan. The slightly U-shaped relationship, that is, the group with high WSC having a small elevated risk of MDE, may reflect a dark side of WSC in a country with collectivity-oriented and hierarchy-oriented culture, such as Japan.