Abstract: According to different operation scenarios and failure states of pipelines, a risk assessment model and a maintenance decision-making model based on Bayesian network (BN) were proposed. The hazard source identification was carried out for the external corrosion of the pipeline, and the external corrosion risk assessment model was established according to the identification results. Then, the existing maintenance methods and corresponding costs of external corrosion of pipelines were summarized and the maintenance decision model was established by combining the statistical results and risk assessment model. The optimization function was created by integrating the maintenance cost and external corrosion loss. The function can be used to develop the maintenance decisions and can realize the reasonable choice of maintenance strategies under the situation of external corrosion of pipelines.