To explore the association between Aerosol Optical Depth(AOD) with air pollutants and meteorological factors, and to provide evidence for the prediction of AOD, time series analysis was used. The data of AOD, air pollutants, and meteorological factors were from January 2002 to December 2004 in Hangzhou. The data of the first 35 months was collected to establish ARIMA（Autoregressive Integrated Moving Average）model, and the data of the last month was used for evaluation. After a time difference, ARIMA(0,0,2)(0,0,0) model was established, the Ljung-Box Q test and the residual plots showed the model fit well. The true values were all above 95% confidence level. From the model, the AOD was not only affected by the air pollutants and meteorological factors on the same day, but also by those in several days before.