Abstract: Risk prediction, prevention and control based on massive operation and maintenance data is the basic task of the railway data center to realize artificial-intelligence-based operation and maintenance. Based on the requirements of intelligent operation and maintenance of the railway data center, four intelligent operation and maintenance data analysis methods are studied. Relying on the big data storage and data sharing service capability of railway data service platform, using data analysis, model training, model deployment and other utilities provided by the platform, risk prediction models of several operation and maintenance scenarios such as capacity trend prediction, log analysis-based risk prediction, operation anomaly prediction and construction risk prediction are established, and model training, tuning and testing are also completed. Finally, the models verified via test are released and updated online. The establishment of the risk prediction, prevention and control system for railway data center based on massive operation and maintenance data can improve the evaluation index and prediction model through the accumulation of experiences in operation and maintenance, improve the accuracy of risk prediction and the effectiveness of risk disposal, and help operation and maintenance personnel quickly focus on major problems, thus guaranteeing the long-term safe and stable operation of the railway data center, and consolidating the foundation of railway transportation production safety.