Due to the defect in a system or design, some elements of one system may often failure unconventionally, which caused by “early failure”. Usually, we eliminate these obvious extraordinary data before analysis them in reliability trials to reduce interfere. However, this approach was so subjectively. Change point models are usually used to study for the time points, which change suddenly in the system models, and are important in many applications. In this paper, we are focusing on finding out those change points that can disturb the evaluation for the reliability model. We bring forward a Bayesian change point model, which is very popular in clinical trials. By this model, we use the time lagged regression function, which is based on the stochastic process, to do the change point analysis for the system reliability data under two hypothesis, by which we can evaluate the covariates’ influence on the life distribution for the system more correctly. What’s more, in this paper we make use of the Markov chain Monte Carlo (MCMC) approach based on Gibbs sampler to simulate dynamically the Markov Chain of the parameters’ posterior distribution. And also, we give out the parameters’ Bayesian estimation in the condition of the random truncated test. Finally, we utilize the result of the data simulation to prove the objectivity and validity of the model by using the WinBUGS package, and the univeriate example cited here could be extended to multivariate data and also be helpful for the study of “ bathtub curve”.