Corrosion is one of the main degradation factors leading to failure of subsea pipelines. The periodic maintenance is an important technical approach to reduce and control the failure risk of corroded subsea pipelines. However, development of maintenance strategy including interval and scheme poses a concern of risk and cost. This paper proposes a risk-based model to develop an optimal maintenance strategy for corroded subsea pipelines using Bayesian network and Markov process approach. When the failure risk is out of the acceptable criteria, the shorter maintenance interval should be implemented. To determine the optimal maintenance interval, the failure scenario of corroded subsea pipelines is developed using coupling Bow-Tie (BT) and Bayesian network (BN). The required value of failure probability is estimated by the probability updating. Subsequently, the optimal maintenance interval is determined by the required probability and failure probability distribution. Besides, the Markov process is employed to estimate the availability and corresponding cost of corroded pipeline under different maintenance schemes. Thus, the optimal maintenance scheme is determined, which can achieve a balance between availability of pipeline and maintenance cost. Eventually, the proposed methodology is demonstrated by a specific case study. In essence, the present study can provide an educational reference for the risk management or decision making of corroded subsea pipelines. (C) 2017 Elsevier Ltd. All rights reserved.