Abstract: Alleviating the pollution of atmospheric particles and creating a clean and healthy urban environment is one of the problems to be solved in most cities in China. Urban green space can effectively reduce the PM2.5. However, the research on the spatial forms of green space is insufficient, which leads to few achievements that can be directly applied in planning and design. Based on the quantitative difference of spatial morphological elements and field measurement of PM2.5 concentration of nine urban green spaces in the city of Wuhan, this study exlpores the deep quantitative relationship between green spaces and PM2.5. The results show that: 1) the decrease of PM2.5 increases with the expansion of the green space scale. The reduction will slow down when the scale is larger than 5~10 hm2; 2) The area, perimeter and shape index of the green space are positively correlated with the reduction of PM2.5. The fractal dimension index and the related circumscribing circle have no significant correlation. 3) The reduction efficiency of PM2.5 decreases sharply with the increase in the green space scale. The reduction efficiency is not obvious when the scale of green space exceeds 5~10 hm2. Through quantitative analysis, this paper reveals the spatial morphology index of urban green space, which has a significant impact on PM2.5 reduction, and puts forward some green space optimization strategies, such as appropriate scale, irregular or complex green space forms and decentralized layout.