Abstract: In order to recognize the pollutant sources and build the correspondence relationships between contaminated sources and important pollutants,a set of intelligent recognizing method based on correspondence factor analysis and fuzzy C-means clustering(short for IRM-CFA&FCM)is developed IRM-CFA&FCM couples skilfully fuzzy C-means clustering,correspon dence factor analysis and clustering validity funct ion discovered by Xie,etc,and is solved by the accelerating genetic algorithm efficiently.The method is used to monitor water quality in Donghu lake,acquiring six kinds of correspondence relationships and puting forward some reasonable and valuable suggestions for decisions making of renovating Donghu's contamination comprehensively.The method is useful for recognizing pollutant sources of other city water bodies.