While previous evidence suggested the regional variation of associations between particulate matter air pollution and cardiovascular disease, the regional characteristics that derive the variation remain uncertain. We aimed to identify regional characteristics which determines the variation of associations of PM10 and cardiovascular disease (CVD) incidence across 16 provincial-level regions in South Korea.We used the National Health Insurance-National Sample cohort 2.0 which is constructed from the National Health Insurance database for the entire South Korean population for 2002-2015. Our study population included 155,017 adults who resided in the same regions for the entire follow-up period. South Korea is comprised of 16 regions: 7 metropolitan cities and 8 provinces (average size: 6,242 km2, [range: 501-19,031]; average population: 2,954,934 [531,887-10,415,399]). The incident CVD was defined as the first diagnosis for 2007-2015, while individual long-term PM10 exposure was assessed as a 5-year average concentration for 2002-2006 predicted at district-level residential addresses. We applied time-dependent Cox’s proportional hazard model to estimate region-specific hazard ratios (HRs) for incident CVD per a 10 μg/m3 increase in PM10 after adjusting for individual characteristics. For regional characteristics, we obtained 76 variables in seven categories: medical accessibility, socio-economic status, behavior, neighborhood facility, social awareness, emission sources, and air pollution and meteorology from various data sources such as census, national survey, and general statistics. To investigate the spatial variation of HR of CVD for PM10, we computed Pearson’s correlation coefficients between HRs and each of regional variables across 16 regions.HRs varied between 0.67 and 1.46 across 16 regions. Region-specific HRs were significantly positively correlated with percent of paved roads, city gas supply, and gross generation from diesel power plants for 2007-2015.Our results suggest that the impact of long-term PM10 exposure on incident CVD could be stronger in the areas with industry-related emission sources.