Abstract: Aiming at the problems with equality constraints for the optimization model of water allocation in the irrigation areas and according to the features of particle swarm optimization, the large scale system model based on the optimal algorithm was developed. Meanwhile an improved particle swarm optimization which can strictly control the speed of each particle was suggested. The deficiency of the model that the speed of each particle is so high that the particle is easy to escape the constraint space causing the interruption of computer programme was overcome. The model was applied to the optimal allocation of water resources in irrigation areas, for example, the limited water resource is allocated to the different subareas, different crops in the subareas and different growth stages of crops. The model is proved to be effective and efficient.