The invention relates to an electric power pollution grade evaluation method of AOD data based on MODIS. AOD sensory data obtained by the MODIS has the advantages of high timeliness, wide coverage andthe like, so that the problem of poor spatial representativeness in a polluted area map making process can be solved, and the workload and cost of monitoring and evaluation can be greatly reduced. Aiming at the problems of the existing pollution evaluation method, the invention provides a novel large-area low-cost electric power pollution evaluation method by utilizing atmospheric aerosol opticalthickness (AOD) data and introducing a data mining and image classification technology.