The work aims to study the whole failure process of submarine pipeline under the dual effects of pitting corrosion and corrosion fatigue, and construct a system failure model based on dynamic Bayesian network to predict the failure probability of submarine pipeline system under different fatigue life. The pitting fatigue damage process was divided into four stages: pit nucleation, pit growth, short and long crack growth. The Monte Carlo simulation method was used to analyze the pipeline failure process from pitting formation to short crack occurrence. Based on the dynamic Bayesian network structure diagram of fatigue crack growth and the uncertainties of related factors, an innovative probability analysis method for submarine pitting pipeline system was proposed to scientifically predict the failure probability of pitting pipeline fatigue life. Combining with the example analysis, the critical crack size of pit growth to short crack growth was 0.8 mm, which was solved by the Monte Carlo simulation method. The fatigue life of pitting pipeline without maintenance was predicted by the dynamic Bayesian network analysis method. The pipeline would face failure risk after 35 years of working. The results show that the model can reasonably predict the failure probability of corrosion-fatigue life of subsea pitting pipelines. By observing the changes of relevant influencing parameters and updating the predicted results in time, it is helpful to formulate effective maintenance strategies for subsea pipeline systems.