Abstract: In order to meet the needs of in-depth analysis of the causes of railway Electric Multiple Unit (EMU) failures and prediction of trend laws, this paper analyzed the current situation of on-board non real-time data, proposed the method of non real-time data download using Wireless Local Area Network (WLAN) and 5G mobile communication system public network, and designed the platform of on-board non real-time data download and application for EMU. Through the key technologies of massive data processing and monitoring, batch efficient data analysis and Data integrity verification, the paper realized the function of the collection and transmission, data monitoring, data parsing, data sharing and comprehensive display of vehicle borne non real-time data. The implementation of this platform can further promote the download construction of non real-time data of EMU, and is of great significance to the research on fault analysis and operation status assessment and prediction of EMU.