Abstract: The accurate establishment of a dynamic model for electric drive gear systems and the analysis of their nonlinear vibration characteristics serve as the core prerequisite for implementing system vibration suppression and performance optimization. In this research, the time-varying mesh stiffness (TVMS) of the gear pair is calculated by the slice method combined with the integral idea, and the time-varying contact force of the bearing is solved based on the Hertz contact theory. and then the nonlinear dynamic model of the gear-bearing coupling is established by comprehensively considering the multi-source excitation factors such as transmission error and backlash. Subsequently, the fourth-order Runge-Kutta numerical integration method is employed to solve the established nonlinear dynamic equations, thereby obtaining the dynamic response of the system, on this basis, the time-domain features, phase trajectory patterns, and bifurcation behaviours of the system are systematically analysed, and the quantitative influence law of excitation frequency changes on the amplitude, waveform, and other characteristics of vibration responses is explored in detail. Furthermore, from a global dynamic perspective, the evolutionary rule of the system’s attraction domains during the variation of different excitation frequencies is investigated, aiming to reveal the multi-stable phenomena and their corresponding stability characteristics under different parameter configurations. Finally, the effectiveness of the proposed dynamic model and the accuracy of its dynamic response prediction are verified through a comprehensive comparative analysis between the experimental data measured on a dedicated electrified drive assembly test rig and the numerical simulation results. The research findings indicate that excitation frequency exerts a significant regulatory effect on the nonlinear vibration characteristics of the electrified drive gear system, and within specific frequency intervals, the system exhibits complex dynamic behaviours such as single-periodic motion, multi-periodic bifurcation, and even chaotic motion. These research achievements provide a solid theoretical support and reliable technical reference for the vibration suppression, stability improvement, and dynamic performance optimization of electrified drive gear systems in practical engineering applications.