Abstract: Irrigation canal water distribution is a problem of space-time canal discharge distribution in irrigation processing. Conventional artificial methods are difficult to optimize the canal discharge distribution, and they easily lead to a great deal abandon water. So, the optimization of canal discharge is an important part of irrigation district modern management, it also has an important role in improving the level of irrigation district management and water saving. In this paper, the authors modify the optimization model of rotation irrigation branch canal distribution in 0—1 linear integral programming. Using genetic algorithm, the model is solved. The genetic arithmetic operators such as tournament selection and arithmetic crossover and exchange mutation are used . Based on these operators, a Matlab 6.0 program is developed to solve constrained 0—1 linear integral programming. The optimization results show that the results are better than conventional models, it indicate that the model and it's algorithm are feasible.