Abstract: The structure and algorithm of fuzzy ARTMAP neural network are described in detail, and an automatically adjusted algorithm of vigilance parameters is presented. The automatic adjustment has the advantages of inducting classification results into the selection of vigilance parameters and avoiding the problem of getting the optimum value artificially. The classification results show that the fuzzy ARTMAP neural network with the automatically adjusted algorithm can effectively classify the thematic mapper(TM) image which covers Xianghai Nature Reserve, and it depends less on samples and gets higher accuracy of the classification than maximum likelihood classifier and conventional fuzzy ARTMAP neural network.