岩溶水库渗漏受众多因素影响 ,对其渗漏量预测一直是水利工作者的研究热点。本文根据非线性理论 ,综合考虑各种地质因素并由此建立了一个神经网络系统。研究结果表明 ,该神经网络模型具有较高的预测精度 ,是可行和实用的。在此基础上 ,利用此网络模型对水布垭水利枢纽渗漏进行了定量预测 The leakage of karst reservoirs is affected by many factors, and the prediction of the leakage of karst reservoirs has always been a hot topic for water conservancy workers. In this paper, based on the nonlinear theory, a comprehensive consideration of various geological factors and the establishment of a neural network system. The results show that the neural network model has high prediction accuracy and is feasible and practical. On this basis, this network model is used to quantitatively predict the seepage of Shuibuya Water Conservancy Project