Lidar plays an important role in monitoring atmospheric aerosol for its high vertical spatial and temporal resolution. During the process of inversing aerosol extinction coefficient and optical depth by Fernald method, Lidar ratio is the main error source. However, in the present application, Lidar ratio is commonly taken as a fixed value, which may inevitably bring a certain amount of errors to the retrieval from Lidar measurement. An estimating model to calculate varied Lidar ratio through the stepwise regression analysis of pollutant data, meteorological data and Lidar data is established. The study shows that Lidar ratio has significant correlation with PM2.5 mass concentration, relative humidity and the mass concentration of SO2 and NO2. Compared to the fixed value, the usage of varied Lidar ratio can significantly improve the inversion accuracy of aerosol extinction coefficient and optical depth.