Background: Particulate pollution is currently regarded as a severe environmental problem, which is intimately linked to reductions in air quality and human health, as well as global climate change. Objective: Accurately identifying the key factors that drive air pollution is of great significance. The temporal and spatial heterogeneity of such factors is seldom taken into account in the existing literature. Method: In this study, we adopted a geographically and temporally weighted regression model (GTWR) to explore the direction and strength of the influences of natural conditions and socioeconomic issues on the occurrence of PM2.5 pollutions in 287 Chinese cities covering the period 1998 to 2015. Result: Cities with serious PM2.5 pollution were discovered to mainly be situated in northern China, whilst cities with less pollution were shown to be located in southern China. Higher temperature and wind speed were found to be able to alleviate air pollution in the country’s southeast, where enhanced precipitation was also shown to reduce PM2.5 concentrations; whilst in southern and central and western regions, precipitation and PM2.5 concentrations were positively correlated. Increased relative humidity was found to reinforce PM2.5 concentration in southwest and northeast China. Furthermore, per capita GDP and population density were shown to intensify PM2.5 concentrations in northwest China, inversely, they imposed a substantial adverse effect on PM2.5 concentration levels in other areas. The amount of urban built-up area was more positively associated with PM2.5 concentration levels in southeastern cities than in other cities in China. Conclusion: PM2.5 concentrations conformed to a series of stages and demonstrated distinct spatial differences in China. The associations between PM2.5 concentration levels and their determinants exhibit obvious spatial heterogeneity. The findings of this paper provide detailed support for regions to formulate targeted emission mitigation policies.