Abstract: Bue to the excessive conservatism and insufficient validity in the traditional methodology of remaining life prediction for corroded pipeline, sources and characteristics of systematic measurement bias and random error in the internal inspection datas were analyzed, and the relevant solution of internal inspection data correction was determined. Based on the theory of normal distribution, the growth rate calculation model was established, and the growth rate corresponding to the 95% upper limit quantile was used to predict the remaining life of the pipeline with defects. This method considered the conservation and rationality of maintenance plan. The residual life prediction results were compared and analyzed with typical engineering cases. It was found that the improved residual life prediction method could be helpful to establish a more scientific and reasonable maintenance response strategy and to improve the efficiency of pipeline operation.