The signal to noise ratio(SNR) of seismic waves is usually very low after long distance transmission. For this condition, to improve the bearing estimation capability in the low SNR,a frequency domain polarization weighted ESPRIT method using a single vector device is proposed. The frequency domain polarization parameters extracted from the signals are used to design the weighted function which is applied to the received signals. The bearing angle and the target frequency are estimated through ESPRIT using the weighted signals. The simulation and experiment results show that the presented method can obtain accurate estimation values under the low SNR with little prior information. The signal to noise ratio (SNR) of seismic waves is usually very low after long distance transmission. For this condition, to improve the bearing estimation capability in the low SNR, a frequency domain polarization weighted ESPRIT method using a single vector device is proposed. The frequency domain polarization parameters extracted from the signals are used to design the weighted function which is applied to the received signals. The bearing angle and the target frequency are estimated through ESPRIT using the weighted signals. The simulation and experiment results show that the presented method can obtain accurate estimation values ​​under the low SNR with little prior information.