Abstract:The Van Genuchten model (hereinafter refer to as VG model) is the most wide use model for soil water characteristic curve at present, put forward the feasible optimization algorithm to identify the model parameters is also a very important research direction. In this paper, pointed to the disadvantage of the standard particle swarm algorithm was easy to fall into local optimum, presented a multiple neighborhood particle swarm algorithm which can be effectively overcome the shortcoming of local optimum, also can use this algorithm to identify the parameters of the VG model, finally different type of soil moisture performances are tested by these parameters. The numerical experiment results showed that: The multiple neighborhood particle swarm algorithm can be effectively applied to identify the parameters of the VG model, compared with other algorithms in terms of performance and precision are improved, also the scope of parameters can be larger. As a result, The multiple neighborhood particle swarm algorithm can be used as a new kind of method for identification of the VG model parameters.