Hydraulic systems are vital in industry, yet accurately modeling them presents significant computational challenges. Traditional methods for adjusting parameters are often inefficient. In this study, we employ a data-driven optimization strategy that utilizes a particle swarm optimization algorithm with surrogate models to effectively identify model parameters, ensuring feasible solutions at reduced computational costs. Experimental validation conducted on a servo valve-controlled hydraulic cylinder simulation model confirms the efficiency and accuracy of our approach in adjusting parameters. This research addresses challenges in hydraulic system model parameter identification by focusing on enhancing computational efficiency and effectiveness. We propose a data-driven optimization approach, which begins by defining an optimization problem based on credibility indicators. Using a Particle Swarm Optimization (PSO) algorithm paired with surrogate models, our method employs dynamic partitioning and parallel optimization to swiftly identify potential solutions within constrained computational costs. Experimental validation on a servo-valve-controlled hydraulic cylinder system confirms the effectiveness of our approach in parameter identification and optimization. This method significantly improves optimization efficiency and holds promise for similar applications in hydraulic systems, contributing to practical solutions in system modeling.