Abstract: Air pollution is an unavoidable problem in the rapid development of urbanization in many countries. The paper takes Wuhan, a city of rapid urbanization, as the object of study. Through remote sensing, it has inversion of the atmospheric aerosol optical depth, interpretation of land cover types and, based on the pixel binary model, inversion of NDVI to obtain the vegetation coverage. Then, it explores the correlation between AOD and vegetation cover on the grid scales of 500m×500m, 1km×1km, 2km×2km, and 3km×3km. The results show that under the four scales, they all have extremely negative correlation at the 0.01 level, with close pearson coefficient. Then, it applies four kinds of grid-scale regression models as vegetation coverage and AOD prediction models, to reveal initial changes of the two. Further analysis at a grid scale of 1 km shows that when the vegetation coverage rate is ≤10% and >45%, there is a significant negative correlation between the two levels at the 0.01 level. The area with vegetation coverage rate ≤10% is highly spatially consistent with the water body and its surrounding areas and urban and rural construction land. The overall pollution is serious. The area with vegetation coverage rate >45% is mainly distributed in forests, open forests, and agricultural land. In ecological land areas where vegetation coverage is good, the pollution is mainly moderate, light, and fine, indicating that increasing vegetation cover is of great significance in alleviating air pollution. In this study, the atmospheric aerosol optical depth comprehensively reflecting air pollutants is obtained by remote sensing inversion, and to inverse the vegetation coverage rate more accurately, the improved pixel bipartite model is used based on NDVI. This paper explores the correlation law and spatial distribution law between AOD and VCR, and provides a quantitative reference for urban and rural planning and design from the perspective of mitigating comprehensive air pollutants.