This study aims to address the issues of high energy consumption and the difficulty in dynamically adjusting equipment parameters in the operation of water systems in refrigeration plants. A modeling and optimization method driven by measured data is proposed. First, the TRNSYS software is used to build an operational model of the refrigeration plant’s water system, incorporating key equipment such as screw-type chillers, chilled water pumps, cooling towers, and the piping network. Then, the TRNSYS model is calibrated and adjusted using experimentally measured flow data from the chilled water system of a coal mine refrigeration plant, ensuring that the simulation accurately reflects real operating conditions. Next, the Particle Swarm Optimization (PSO) algorithm is employed to iteratively search and optimize key operational parameters of the system—including chilled water supply temperature, the number of operating chilled water pumps, and pump speed ratios—with the goal of minimizing energy consumption and optimizing system efficiency. The results show that after applying the PSO algorithm, the overall energy consumption of the system across different load ranges was reduced by an average of 13.97%, demonstrating the effectiveness of the proposed method. The conclusion indicates that this method not only significantly reduces the operational energy consumption of the refrigeration plant’s water system and enhances system efficiency, but also provides a feasible technical approach for energy efficiency optimization of similar systems, holding great significance for energy conservation, emission reduction, and sustainable development．