作者采用具有高度非线性识别能力的人工神经网络与遗传算法相结合的方法，探讨了建立智能预报模型的基本方法，进一步对如何提高预报精度的问题进行了研究，并结合黄河洪水预报实例检验了神经网络模型的可行性。-The author used to identify the capacity of highly non-linear artificial neural network combining with genetic algorithm approach to explore the establishment of intelligent prediction model of the basic ways to further improve the forecast accuracy of how the issues of research and examples of combination of the Yellow River flood forecasting tests the feasibility of neural network model.