The corrosion evolution behavior of oil and gastransportation pipeline is complicated, and sufficient data on corrosion influencing factors is difficult to obtain during actual operation. Additionally, the traditional empirical model exhibit significant errors in the long-term predictions. In order to more comprehensively characterize the dynamic characteristics of memory effect and measurement random error of pipeline corrosion Process, and accurately predict the corrosion depth of pipeline inner wall, a non-Markov Wiener Process prediction model is proposed considering the dual influence of measurement error and memory effect. The unknown parameters of the model are estimated and updated by maximum likelihood estimation and Bayesian inference. Based on the theory of weak convergence and the definition of first reach failure time, the approximate analytical formula of pipeline corrosion depth distribution is derived to achieve the prediction of pipeline corrosion depth. Finally, the corrosion monitoring data of the inner wall of Tiangao Line B section in Chongqing Gas Mine is taken as an example to verify theeffectiveness of the method.