The invention provides a water resource optimal allocation method based on a genetic algorithm, and the method achieves the normalization of an economic target, a social target and an ecological target, enables all targets to be subjected to weighted calculation under the same dimensional condition, and enables the weights of all targets to have more practical significance after normalization. Meanwhile, the constructed water resource optimization model has the characteristics of multiple variables, multiple targets, multiple constraints and the like, when the model is solved, a genetic algorithm is adopted, and compared with a traditional solving method, the provided method has great advantages in the aspects of calculation speed, precision, convergence, adaptability and mobility.