Despite the rapid development of cement industry, the production control of cement vertical mills continues to heavily rely on traditional expert experience-based methods. In order to enhance production efficiency, this paper meticulously analyzes the dynamic structure and production processes inherent in cement vertical mill systems. It divides the process from the furnace to the mill outlet into three distinct subsystems and carefully selects mathematical process models to construct a comprehensive system model. Field operation data from the Tangshan Plant of Tianjin Cement Research Institute serve as the basis for identifying model parameters. The paper utilizes the particle swarm optimization algorithm to optimize and determine the most appropriate model parameters for the vertical mill system, ultimately achieving an impressive fitting degree of approximately 90% for the optimized model.