Wind turbine gearbox (WTGB) serves as the core component to transmit power and motion. Throughout prolonged operation extending up to 25 years under harsh conditions, the cumulative effects of fatigue damage and tooth surface wear gradually alter gear dynamic parameters, ultimately impacting overall efficiency. The real-time evaluation of dynamic characteristics and damage accumulation for WTGBs continue to present substantial challenges. In this work, a parameter identification approach based on torsional vibration model and time-domain signals is developed. The gear health status is incorporated with tooth spalling characteristic. Dynamic response and fatigue damage of wind turbine gear transmission are analyzed through the established model. The results indicate that the proposed method can effectively predict gear fatigue damage based on the identified gear meshing stiffness and rotational inertia.