Internal-meshing power gear honing is a crucial finishing process that enhances gear tooth surface texture and minimizes transmission noise. However, chatter is a common issue during honing, significantly affecting machining accuracy and productivity. A precise understanding of the dynamics of gear honing machines is essential for mitigating chatter and improving manufacturing precision. This paper utilizes honing force as the excitation source, integrates machine tool vibration data, and employs the Particle Swarm Optimization (PSO) algorithm for dynamic parameter identification. A new method is presented to establish the dynamics model of a gear honing machine in operational conditions. Experimental validation demonstrates the accuracy of the dynamics model, revealing that the vibration signals from the model closely match those of the actual machine tool, with a root mean square (RMS) error in amplitude within 10%. This modeling approach allows for the establishment of machine dynamics without requiring special designs, operations, or destructive tests, offering a practical solution for real-world production challenges.