Power cable is more and more widely used in power supply and auxiliary systems, such as power, building, communication, etc. However, under the conditions of high voltage, strong electric field, large current and complex environment, power cable is prone to insulation aging, electrical treeing and even breakdown fault. In addition, the heating of copper loss and dielectric loss will lead to the increase and deterioration of cable temperature. Therefore, monitoring cable insulation by monitoring cable partial discharge and temperature is an important means to ensure the safe operation of power grid. In this paper, a high frequency detection module (100MHz) with high-speed acquisition function is developed. First of all, based on this module, combined with synchronous signal noise reduction technology, this paper filters out all kinds of interference signals on site; combined with clustering multi-source separation technology, it solves the problem of multi-source superposition impact type diagnosis; combined with wavelet filtering noise reduction technology, it effectively filters out narrow-band interference signals. In addition, based on a large number of on-site practical experience, a fingerprint database of discharge information including various typical discharge types in each development stage is constructed to realize the accurate diagnosis of PD defect types and the state evaluation and early warning based on the discharge development stage. Finally, an improved method of low-cost high-frequency PD and temperature detection is proposed, which realizes the simultaneous monitoring function of PD and temperature, and improves the reliability of PD and multi-dimensional temperature detection of power cable.