We carried out a longitudinal study on the associations between residential greenness and depression risk in urban areas in Finland. Residential greenness indicators were estimated within various buffer sizes around individuals' home locations (selected n = 14424) using time-series of normalized differential vegetation index (NDVI) and CORINE land cover data (CLC). We estimated individuals' cumulative exposure to residential greenness over a 5-years and 14-years follow-up. We used doctor-diagnosed depression and Beck Depression Inventory for depression assessment. Our multi-logistic model showed an inverse association between residential greenness and depression, implying lowered depression risk for individuals with higher residential greenness. The association was particularly evident when using NDVI-based residential greenness (within a buffer of 100 m radius) and doctor-diagnosis depression data, adjusted with individual-level covariates. The odds ratio was 0.56 (95% CI 0.33 to 0.96) for the 5-years follow-up, and 0.54 (95% CI 0.30 to 0.98) for the 14-years follow-up. The associations between CLC-based total residential green space and depression varied across the different buffer sizes. In general, all the associations depended on the type of depression assessment, quality of greenness indicators, and the spatial scale of analysis. The associations also varied across the socio-demographic groups and neighborhood socioeconomic disadvantage level.