<p>The dynamic characteristics of the electric drive transmission system are complex, with vibration characteristics impacting the stability and lifespan of the mechanical system. Therefore, it is essential to analyze dynamic behavior to enhance system performance. To address this challenge, constructing an accurate and quantifiable dynamic model is a priority. In this paper, the EDTS dynamic model is developed by utilizing the lumped mass method, with bearing contact loads integrated into the formulation. Furthermore, the model’s feasibility is validated through simulations and experiments. The effects of rotational speed, rotor eccentricity, and bearing wear on the time-domain and frequency-domain vibration characteristics of the EDTS rotor system in the x-direction are analyzed. By simplifying the developed centralized mass model, a three-inertia modeling technique is introduced. Additionally, a sliding mode control strategy incorporating RBF neural network compensation is proposed to control the three-inertia model of the EDTS. The neural network effectively identifies uncertain components and enhances tracking accuracy. The control simulation is carried out. The simulation results demonstrate that the proposed control approach is effective in mitigating torsional vibrations and exhibits robustness across diverse structural parameters. The average absolute error of NNSMC strategy is 16.67 % lower than that of SMC strategy.</p>