When the train is running, the strong moving noise is propagated to the underground space, forming a kind of green environmental protection and strong energy artificial seismic wave. Due to the fixed shape of the load system, regular train running time and high repeatability of the source, this kind of train source is widely used in track safety monitoring, near-surface imaging and urban underground space detection. Distributed Acoustic Sensing (DAS) is a new seismic acquisition technology, which has the advantages of dense sampling, simple layout and strong anti-electromagnetic interference ability. In recent years, it has been gradually applied in the fields of microseismic monitoring of urban traffic events, crack detection and underground space imaging. Based on the data of train moving noise in Qinhuangdao test ground obtained by DAS technology, 30 train traffic events were picked up to carry out the attribute analysis of train noise's apparent velocity, waveform, spectrum and energy changes in time and space. The results show that the optical fiber sensing train noise in the test area has obvious attribute difference. The seismic amplitude and energy caused by the train with low speed and heavy load are larger than those caused by the high-speed train, and the noise produced by low-speed trains contains rich low-frequency information. According to the known apparent speed, amplitude and time-frequency attributes, four types of trains are divided, and their rules of entering and leaving the station are revealed. This study expands the application scenario of DAS and proves that DAS is expected to be applied to commute monitoring of different types of train.