The evaluation index system and grading standards of regional water environment carrying capacity were built by using analytic hierarchy process. Based on the algorithm principle of Elman neural network and generalized regression neural network ( GRNN ), the paper proposed the evaluation model of water environment carrying capacity of Elman and GRNN neural network, used the interpolation method to construct network training samples, reviewed the threshold sample of water environment carrying capacity evaluation standard, took the results as the division basis to evaluate the carrying capacity class of a regional water environment, and evaluated the water environment carrying capacity of Wenshan in different level years. The results show that the water environment carrying capacity of Wenshan state in different level years is between absolute bearing and basic bearing, witch objectively reflects the present situation and plan desired effect of regional water environment, and can provide reference for the evaluation and research of regional water environment carrying capacity. The evaluation results of Elman and GRNN neural network models are similar witch shows that the model and method of evaluation of regional water environment carrying capacity established by study are reasonable and feasible, the two can be used as the evaluation model of regional water environment carrying capacity.