In order to evaluate the surface water environment quality accurately and objectively,the BP neural network with improved algorithms is introduced in water environmental quality evaluation.Water environment quality evaluation model of the Weihe River was constructed with recorded data of water quality during the period from Sep.7 to Oct.11 in 2010.Water quality of the Weihe River was evaluated through 3 water quality monitoring projects(NH₃N,potassium permanganate index,DO),and the evaluation results by the BP neural network and by the environmental protection department were compared,and both evaluation results are identical.The results showed that the BP neural network model could well solve complex nonlinear relation between the evaluation factors and water quality levels,and the method was simple and reliable,and the forecast precision was high.The BP neural net work for water environmental quality evaluation is objective,universal and practical.