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.