Abstract: This paper firstly extracts the impervious surface data of Beijing in four periods of each year from remote sensing image data based on the remote sensing platform and inverts the AOD value. Then, using the idea of M-estimated robust regression, regression analysis was performed on the AOD value and the monitoring site PM_2.5 data, and a regression model was established. Based on the regression model and the inverted AOD data, spatially continuous PM_2.5 concentration data was generated. Finally, this paper discusses the mechanism of urban sprawl's influence on the spatial and temporal distribution and evolution of PM_2.5 pollution, and quantitatively analyzes the relationship between them. The results show: In 2000, 2006, 2012 and 2016, the impervious surface areas of Beijing are respectively 6 646.37 km², 9 680.52 km², 9 736.31 km² and 9 769.20 km². Within 2000 and 2016, the growth rate of impervious surface was 46.99%, and the corresponding growth rate of PM_2.5 concentration was 56.61%. In the same period, the PM_2.5 pollution in Beijing was first aggravated and then decreased fluctuantly. Spatially, there was a gradually increasing tendency from the northwest direction to the southeast and the serious polluted areas were the east of Fangshan District, Daxing District, Tongzhou District, Haidian District, Chaoyang District, Fengtai District, Shijingshan District, Dongcheng District, Xicheng Districtt, Xuanwu District and Chongwen District. During that period, the spatial distribution of impervious surface in Beijing was highly consistent with the spatial distribution of PM_2.5 pollution. The impervious surface of Tongzhou District, Shunyi District, Pinggu District and Daxing District in the southeast of Beijing was more than 90%, and the growth of PM_2.5 concentration in these regions was also higher than that in the northwest.