The temporal variation of aerosol optical properties retrieved from the Aerosol Robotic Network（AERONET） at the Taihu site in East China is analyzed. The classification study of aerosols is conducted by the method of cluster analysis based on the Mahalanobis distance. The results show that the mean values of aerosol optical depth (AOD) reached a maximum of 1.29 in summer and the maximum value of Angstrom Exponent (AE) is 1.31 in autumn. The highest average value (0.92) of single scattering albedo (SSA) was shown in spring and autumn, while the maximum average value (0.66) of asymmetry factor appeared in summer. The overall seasonal volume size distribution of aerosols showed a bimodal model，with a fine mode peak at about 0.26 μm in summer and a coarse mode peak at about 2.24 μm in spring. The coarse mode aerosols were mainly from the transport of dust aerosols from the north part of China and the marine aerosols from sea. The fine mode aerosols were mainly caused by biomass burning mixed with urban industrial pollutants. The maximum of absorbing aerosol optical depth (AAOD) in summer is 0.11, while the minimum value of AE is 1.0, indicating that aerosol particles in this area contain typical black carbon aerosols. Mixed aerosols accounted for the largest proportion (54.09%) in the Taihu Lake region, followed by urban industrial aerosols (18.95%). Seasonal differences in the proportions of the aerosols of the five categories were more pronounced, and mixed aerosols accounted for the largest proportion in autumn (65.30%). The urban industrial type accounted for the highest proportion in summer (32.07%), and in the spring due to the transport of dusts and aerosol particles in the north, the proportion is the highest (21.01%).