Abstract: Background: Residing in greener areas may decrease the burden of chronic diseases, but the association with cancer is unclear. We studied the associations between residential green spaces and site-specific cancer mortality in urban Belgium. Methodology: We linked the 2001 Belgian census, register mortality data for 2001-2014, and environmental information (green spaces and air pollution) at baseline residence (2001). We included residents from the largest Belgian urban areas aged >= 30 years at baseline. Exposure to residential green spaces was assessed using the Normalized Difference Vegetation Index (NDVI), Urban Atlas, and perceived neighbourhood greenness (from the census). We used Cox proportional hazards models to obtain hazard ratios (HR) and their 95 % confidence intervals (95 %CI) of the mortality risk from lung, colorectal, breast (in women) and prostate cancer (in men) per interquartile range increment in residential green spaces. We further analyzed the role of outdoor air pollution and effect modification by age and socioeconomic position (SEP) in main associations. Results: 2,441,566 individuals were included at baseline. During follow-up, 1.2 % died from lung cancer, 0.6 % from colorectal cancer, 0.8 % from breast cancer, and 0.6 % from prostate cancer. After adjustment, higher exposure to green spaces was associated with a reduced mortality risk from lung cancer and breast cancer [e.g., for NDVI within 300 m, HR:0.946 (95 %CI:0.924,0.970), and HR:0.927 (95 %CI:0.892,0.963), respectively], but not with colorectal or prostate cancer mortality. For the latter, a suggestive hazardous effect of green spaces was found. Air pollution seemed to have only a marginal role. Beneficial effects of greenspace were generally stronger in < 65-year-old, but no clear trend by SEP was found. Conclusions: Our findings suggest that residing in green areas could decrease mortality risk from lung and breast cancer, potentially independent from air pollution. Future studies should consider different ...