In this paper, we present the new particle swarm optimization algorithm by defining the inertia weight which depends on linearly the iteration step for solving the unsaturated soil water problem. By comparing with the exist particle swarm optimizations such as such as PSO algorithm, FPSO algorithm, P-PSO-SA algorithm and IDWPSO algorithm, numerical experiments verify that our proposed algorithm is more efficient and accurate to reach the optimal solution in the multimodal function extremum problem. Combining the characteristic difference method, our algorithm is applied to solve the unsaturated soil water problem. Numerical experiments show that our algorithm is quicker and more precision to approximate the exact solution. Keywords: Particle swarm optimization; Unsaturated soil water; Inertia weight; Characteristic difference; Parameter identification