The recent increase of railway traffic volume and improvement in train operation speed has lead to a rise in the national awareness of and public demand for railway safety. As a result, the establishment of safety measures is urgently required. Together with the importance of railway as a key form of infrastructure and the most efficient means for human and material exchange to help determine the competitiveness and growth of national industry, the necessity to secure railway safety is being emphasized. Accordingly, with a goal to prevent railway accidents, this study was conducted to develop a system that extracts safety-related data, such as information about rolling stock failure and the conditions of tracks, power facilities and structures that can lead to accidents, establishes big data on railway safety using the extracted data, provides data modeling and accident prediction results in real-time to railway engineers and controllers, and thus improves decision-making ability for appropriate accident risk control and accident handling.