本文在用人工神经网络BP模型对流域年均含沙量进行多因素建模过程中 ,对BP算法进行了改进。在学习速率 η的选取上引进了一维搜索法 ,解决了人工输入 η时 ,若 η值过小 ,收敛速度太慢 ,η值过大 ,又会使误差函数值振荡 ,导致算法不收敛的问题。建模实践表明 ,改进后的BP算法可能使网络误差函数达到局部极小点 ,提高了算法的拟合精度 In this paper, artificial neural network BP model for annual sediment concentration in the multi-factor modeling process, the BP algorithm is improved. In the selection of learning rate η, a one-dimensional search method is introduced to solve the problem that if the value of η is too small, the convergence speed is too slow and the value of η is too large, the value of the error function will oscillate, resulting in the non-convergence of the algorithm problem. Modeling practice shows that the improved BP algorithm may make the network error function reach a local minimum and improve the fitting accuracy of the algorithm