Rotary lip seal is widely used in aircraft and its performance affects the safety of the aircraft. Hence, it is necessary to estimate useful lifetime and reliability of the seal. Degradation of rotary lip seal is always with random effects, which cannot be considered by theoretical failure mechanism analysis. Hence, in order to consider the random effects of rotary lip seal degradation, stochastic processes are applied. Furthermore, considering the monotonic degradation of the seal, Gamma process and inverse Gaussian process are selected as the candidate processes. To combine the candidate processes, Bayesian model averaging is introduced. Based on the failure mechanism analysis and numerical simulation, the theoretical wear path is predicted and corresponding linearization method is proposed. The measured degradation data is converted and the seal wear process is transformed to a linear degradation process. The model parameters and model probabilities are evaluated by fully Bayesian inference method. The effectiveness of the proposed method is verified by comparing the predicting degradation and experimental observations. The proposed method can be used to evaluate reliability and useful lifetime of rotary lip seal. According to sensitivity analysis, an effective way to improve lifetime and reliability of the seal is to increase the wear depth threshold.