Identification and detection of faults traveling wave are very important for power grid faults location. A novel time-frequency analysis method for the traveling wave signals is proposed. Hilbert-Huang transform（HHT） is applied to detect the signals. Intrinsic mode functions（IMF） of traveling wave are extracted by the empirical mode decomposition（EMD）， and then Hilbert transform is used to calculate the corresponding instantaneous frequency of each IMF component. The arrival time of traveling wave can be detected by the instantaneous frequency. Compared with the wavelet transform， it is not necessary to select parameters in HHT， and the transform has only one result. Simulation results show that the HHT is more exact to extract the arrival time of traveling wave， and the precision of fault location is in range of practical acceptance.