Abstract: Concrete dam parameter identification is the key to evaluate its behavior. To further improve the efficiency and accuracy of parameter inversion, a novel inversion strategy for concrete dam parameters based on multi-output twin support vector regression (MTSVR) and improved snow ablation optimization (ISAO) is proposed. The nonlinear relationship between the parameters to be inverted and the hydrostatic pressure component of displacement is simulated by training the MTSVR model in place of complex finite element calculations. ISAO is used for optimization inversion of target parameters. The analysis of engineering examples shows that the results of the surrogate model are basically consistent with those of the finite element calculations, and ISAO has faster convergence speed and higher accuracy than the traditional meta-inspired optimization, and the single parameter inversion takes less time. The results show that the constructed inversion strategy can effectively improve the computational efficiency under the premise of maintaining the computational accuracy, and the method has the effectiveness and practicability, which provides a reference for the real parameter identification of engineering.