To solve the problem of slow convergence speed and easy falling into local optimum for general optimization algorithms during parameter inversion of constitutive parameters of earth-rock dams, the adaptive chaotic mutation particle swarm optimization(ACMPSO)algorithm was improved by using dynamic variation coefficient and OpenMP parallel instructions, and the improved ACMPSO algorithm was validated by a case. The results demonstrate that the improved ACMPSO parallel algorithm can effectively prevent algorithm from falling into local optimum and reduce the calculation time significantly. The algorithm has the characteristics of fast convergence, high inversion accuracy and good stability, which is suitable to solve complex parameter inversion problems with high dimensions and huge computation.