Abstract: In this article, many factors about irrigation model were considered, including irrigation water volume, water demand of the crop, planting structure production function of irrigation water, rainfall, soil water balance, the sensitive index of the water depletion to production at different growth stages of crops, the grain market price, irrigation water price, minimum yield, irrigation cost, etc. Then a multi-crop, multi-constraint and non-linear optimization irrigation model based on the maximal profit to irrigation water volume was set up, which adapts to Chinese national conditions, and the real number encoding space of the model was searched by the strong searching ability of genetic algorithm. The results show that this model can solve the optimization irrigation problem of multi-crops, and genetic algorithm has perfect searching function, and can find the optimal solution of the model for a very short time.