Led by certain European countries, rolling stock (railway car) maintenance technology is undergoing a paradigm shift from preventive maintenance based on the inspection period to predictive maintenance in a bid to reduce damages to railroad components that cause interruptions to railroadoperation and incur unnecessary maintenance costs. This has led to increasing demand for fault diagnosis and remaining- useful-life-prognosis technologies in order to simultaneously satisfy the need for greater reliability and lower maintenance costs to cope with faster systems. This study aimed to design the functions and architecture of a rolling stock maintenance support system that analyzes the status data collected from sensors installed at onboard and wayside in order to automatically evaluate and prognosticate the likelihood of parts failures, so as to manage railroad car parts more efficiently.