The invention provides a multi-source data fusion and environmental pollution source and pollutant distribution analysis method. Specifically, the method is mainly composed of data preprocessing, aerosol retrieval, and PM 2.5 inversion; the data preprocessing performs a series of processing on original remote sensing data, outputs data that meets requirements of model operation to be utilized as adata basis of an inversion module, and performs preprocessing operation on the atmospheric remote sensing data used by remote sensing monitoring, and a process of the data preprocessing involves a geometric correction and radiation correction method, a high-precision image registration method, a multi-source satellite data fusion method, cloud removal and shadow processing, and an atmospheric correction link; and the method utilizes GF series data as a main remote sensing data source, combines ground monitoring equipment, develops remote sensing monitoring of particulate matter PM 2.5 and draws a thematic map, and provides remote sensing technology and information support for air quality monitoring and management work of Beijing-Tianjin-Hebei regions.