The problem of high-speed train location in 5G network had been studied by using specific radio (NR) synchronization signal. The path loss, shadow and fast fading effects of 3GPP specified radio channel model are simulated. Time of Arrival (TOA) and Angle of Departure (AOD) are measured from the beam-forming NR synchronous signal. The extended Kalman filter (EKF) is used to track the train position according to the given measured values and the assumed train motion model. The results show that to measurement can achieve better accuracy than AOD measurement, and the best tracking performance can be obtained by considering two measurements at the same time. This method can achieve submicron tracking accuracy in most of the tracking time, thus achieving the expected positioning accuracy requirements of 5GNR.