Aiming for the various type hydropower stations, artificial neural network with Genetic A1- gorithm have been adopted to establish the energy characteristic model of the hydropower station.In this paper, the polynomial method is adopted uniformly to calculate sectional head loss of water diver—sion pipeline for each water turbine generating set respectively.For the relationship of the downstream water level and the flow, besides the application of the polynomial calculation method, for the special case, the logic condition input mode is introduced to be programmed and the system for general optimization of hydropower station is also designed having been in trial operation for some hydropower sta—tions which show the good adaptability in processing the parameters, head loss and downstream water level and so on.