Abstract: The time-frequency domain reflectometry based on pseudo Wigner-Ville distribution was used to detect the time-frequency signal changes of superconducting cable under different temperatures and types of simulated defects. The effectiveness of time-frequency analysis in fault location was analyzed for superconducting cable at different temperatures. The influence of incident waveform shape on fault location was studied by changing the center frequency and bandwidth of incident wave. The domestic 35kV cold insulated three-phase high temperature superconducting cable was used as the simulation sample cable. Three types of faults were set, i.e., loss of insulation, gradual decrease of insulation resistance to ground and short circuit. The detection temperature was set as room temperature, liquid nitrogen temperature and temperature recovery, respectively. The results show that the response sensitivity of this method to the above three kinds of fault increases in turn at room temperature, corresponding location error less than 3%. The time compensation needed for positioning increases with the increase of the center frequency or bandwidth of the incident wave. The peak value of time-frequency cross-correlation at the fault decreases gradually with the decreases of temperature for the short circuit in liquid nitrogen environment. The wave velocity increases obviously when the temperature drops to about 77K. During the temperature rising process, the positioning accuracy of this method is not affected by the temperature change or the cable state. The location error is still below 3%. The change of temperature during the rewarming process can cause obvious synchronous change of the peak value at the fault and the end of the superconducting cable. It can be used to evaluate the temperature rise caused by the leakage of liquid nitrogen.