Abstract: Air quality is closely related to human health and sustainable social and economic development. It is of important theoretical and empirical significance to explore the factors affecting the changes in air pollutant concentration in the context of different land use patterns. This study took the middle reaches of the Yangtze River as an example. We adopted a spatial regression model, in which we used forestland coverage as the benchmark factor, and took the forestland landscape pattern as the key factors to explore the driving mechanism of PM_2.5. Integrated with the regional socio-economic and natural conditions, we selected influencing factors from the three aspects of PM_2.5 namely source, diffusion and removal. The comprehensive influencing factors of PM_2.5 with different forest coverage were explored, compared and analyzed. The results show that: (1) The influencing factors with different forest coverage affect PM_2.5 in similar patterns, yet with different path and magnitude; (2) The spatial heterogeneity of PM_2.5 is obvious. With the increase of forest coverage, the PM_2.5 concentration is greatly reduced; (3) Decreasing the of the fragmentation degree of forest patches is beneficial to reduce the PM_2.5 concentration; (4) DEM is the main factor affecting PM_2.5. It is shown that with the increase of forest coverage, the weakening effect on PM_2.5 increase. Based on the results, it is recommended to strengthen the refined control of air pollution in the middle reaches of the Yangtze River basin according to the characteristics of different forest coverage areas, and combine the spatial spillover characteristics of PM_2.5 to strengthen joint prevention and control among the regions.