建立了以模糊优选、BP神经网络及遗传算法有机结合的智能预报模式与方法。在应用该方法进行中长期水文智能预报时，首先选取训练样本的数量，根据预报因子与预报对象的相关关系得到相对隶属度矩阵;再将其作为BP神经网络输入值以训练连接权重；最后将得到的连接权重值用于预报检验。计算结果表明，智能预报模式与方法的运行速度、精度及稳定性都达到了实际应用的要求。

Intelligent calculating tools such as fuzzy optimization approaches, BP neural network and genetic algorithm are proven to be efficient when applied individually to a variety of problems. Recently，there has been a growing interest in combing all these approaches, and then, in this paper, the author organically synthesizes fuzzy optimal selection, BP neural network and genetic algorithm and establishes intelligent forecasting mode and method. When illustrating the method by an application to forecast mid and long term hydrological process of Yamadu Hydrographic Station at Yili River in Xinjiang, China, the author first selects the amount of training samples, and gets relative membership degree matrix according to the correlation of forecasting factors and forecasting objective, then takes the matrix as input of BP neural network to train link-weights, and finally, uses gained link-weight values to verify forecasting. The results are highly promising and show that the operation speed, precision and stability of intelligent forecasting mode presented in this paper can completely meet actual requirement.