Based on the PM_2.5 monitoring data of China's cities, we identified the spatial and temporal distribution characteristics of PM_2.5 concentrations, and used the geographically weighted regression (GWR) model to analyze emphatically the spatial heterogeneity of the influence of natural factors and socio-economic factors on PM_2.5 concentrations. The results showed that: in 2015, the average annual concentrations of PM_2.5 in China was 50.3 μg/m ³, and the monthly concentration change presented a "U-shaped" pattern with a higher level in autumn and winter while a lower one in spring and summer. In addition, PM_2.5 concentrations were high in cities of eastern and northern China, but low in cities of southern and western China. Beijing-Tianjin-Hebei urban agglomeration was the center of PM_2.5 pollutions in China. The results of geographically weighted regression showed that: (1) in terms of natural factors, elevation had a negative correlation with the urban PM_2.5 concentrations, while positive and negative correlations exist for other indexes, and negative correlation effect dominated, which is conducive to reducing PM_2.5 concentrations in most cities. Thus it can be seen that the influence indexes of PM_2.5 concentrations have significant spatial difference characteristics. From the mean contribution of the regression coefficient, the ranking of the influence intensity of natural indexes on PM_2.5 concentrations were: digital elevation model, relative humidity, temperature, rainfall, wind speed, normalized difference vegetation index. (2) In terms of socio-economic factors, all the indicators showed positive and negative effects, with significant spatial heterogeneity. Among them, the build-up and GDP per capita were conducive to reducing PM_2.5 concentrations in most cities, while population density, foreign direct investment, industrial structure and research and development expenditure can aggravate the air pollution in regions. The ranking of the influence intensity of socio-economic factors on PM_2.5 concentrations were: population density, research and development expenditure, built-up, industrial structure, foreign direct investment, GDP per capita. (3) Due to the spatial heterogeneity of the influence of various factors on urban PM_2.5 concentrations, the spatial difference of the influence of various indexes can be taken into account in the formulation of atmospheric governance countermeasures. Moreover, although natural factors have a more significant influence on PM_2.5 concentrations, since it is difficult to change the natural conditions of cities artificially, specific strategies should be proposed from the perspective of social and economic factors in tackling haze.