While numerous studies have explored the spatial patterns and underlying causes of PM_2.5 at the urban scale, little attention has been paid to the spatial heterogeneity affecting PM_2.5 factors. In order to enrich this research field, we collected PM_2.5 monitoring data from 367 cities across China in 2016 and combined inverse distance weighted interpolation (IDW) and geographically weighted regression (GWR) model. As a result, we could dynamically describe the spatial distribution pattern of urban PM_2.5 at monthly, seasonal, and annual scales and investigate the spatial heterogeneity of the influential factors on urban PM_2.5. Furthermore, in order to make the result more scientific and reasonable, the paper used selection.gwr function and bw.gwr function, respectively, to optimize model, thereby avoiding local collinearity caused by independent variables. The main results are as follows: (1) PM_2.5 in Chinese cities is characterized as time-space non-equilibrium pattern. The Beijing-Tianjin-Hebei region, the Yangtze River corner region, the Pearl River Delta region, and the northeast region have formed a pollution-concentrating core area with Beijing-Tianjin-Hebei region as the axis, which brings greater difficulties and challenges to PM_2.5 governance. (2) The effects of various factors of socio-economic activities on the concentration of PM_2.5 have significant spatial heterogeneity among Chinese cities. (3) There is an inverted “U” curve between economic growth and PM_2.5. When the per capita income reaches 47,000 yuan, the PM_2.5 emission reaches the peak, which proves the existence of environmental Kuznets curve (EKC). These findings could provide a significant reference for policy makers in China to facilitate targeted and differentiated regional PM_2.5 governance measures.