Abstract: A reliable finite element model (FEM) is crucial for the health monitoring and performance evaluation of sluices on soft foundations. Due to the parameter uncertainty, FEM is difficult to accurately reflect the true dynamic characteristics of the sluice. In this paper, we propose an FEM parameter updating method for sluices on soft foundations, which combines modal parameters and is based on the beetle antennae search particle swarm optimization (BAS-PSO) algorithm. The elastic modulus and density that have a great impact on the modal parameters of the sluice are selected as the parameters to be updated. A genetic algorithm support vector regression (GA-SVR) proxy model, which can reflect the nonlinear relationship between the updated parameters and the modal parameters of the sluice on the soft foundation, is established. An objective function based on the minimum relative deviation between the modal parameters identified by the measured response and the modal parameters calculated by the GA-SVR proxy model is proposed. The optimized mathematical model to update the FEM parameter of the sluice on the soft foundation is constructed. A BAS-PSO optimization algorithm is proposed to solve the optimization mathematical model, which overcomes the drawbacks of local optimization and slow convergence. The physical model of the sluice on the soft foundation shows that the modal parameters calculated by the modified FEM are numerically consistent with the vibration modal parameters of the sluice, and that the proposed method is reliable and feasible, providing a new idea for the FEM parameter updating of sluices on soft foundations.