Abstract: Big data provides an effective way for analysis massive historical data in the field of railway safety management to find the regularity of accidents. This article analyzed the present situation of Railway Safety Management System and data, combined with the characteristics of big data, was in accordance with the “people-equipment-environment” three aspects, gave a requirements analysis model for railway safety management of big data applications. This model is an effective way to implement the risk analysis and maintenance of the equipment, provide auxiliary decision-making for decision makers, improve railway safety warning level, which has important guiding significance.