Present paper considers the Bayesian analysis of autoregressive time series model to identify the outlier for a trend stationary time series. We have obtained the posterior probability under different setups of AR (1) time series with linear time trend such as when series is contaminated by additive outlier, when series is trend stationary, and when series is difference stationary. A simulation study has been carried out and it has been observed that the ignorance of outlier has serious impact on stationarity and attests that existence of outlier overestimate the autoregressive coefficients if outlier is not under contemplate.