Accurate reliability assessment of wind turbine gears is important for reducing maintenance costs and improving the sustainability of wind turbine gearbox operation. This study proposes a reliability modeling framework centered on shock–degradation dependence and a time-varying degradation failure threshold. First, a mapping relationship between shock signals and stress is established. Subsequently, a state-dependent shock-induced degradation increment model is introduced to describe the bidirectional mechanism by which shocks accelerate degradation and degradation states amplify shock-induced damage. Furthermore, a time-varying degradation failure threshold model is developed to characterize the dynamic attenuation of the gear load-bearing boundary during service. On this basis, a hybrid Copula function is employed to establish the joint reliability model. A numerical case study was conducted on the sun gear of a 2 MW wind turbine gearbox. The results show that, under the specified numerical operating conditions and model parameter settings, the proposed model under the fixed-threshold condition predicts a relative increase of approximately 6.1% in failure probability compared with the model assuming independent shock and degradation processes. When the time-varying degradation failure threshold is considered, the failure probability shows a relative increase of approximately 30% compared with that under the fixed-threshold condition. The proposed model can effectively characterize the failure evolution of wind turbine gears under complex loading conditions and provide theoretical support for predictive maintenance.