Abstract: Data preprocessing is a preliminary work before data modeling. It can filter the noise interference of original data and can reduce data dimension to extract features of the data in time domain. The large amount of data accumulated in the production process of railway transport industry often contains noise interference. Besides, it is often massive and high dimensional, which cannot be directly used for data modeling, analysis and mining. Principal component analysis and singular value decomposition are important matrix decomposition technologies in linear algebra. They have been commonly used in data time-domain preprocessing in recent years. This paper mainly discussed the use of principal component analysis and singular value decomposition technologies in railway data preprocessing of the application.