In response to the inability of visual observation, infrared temperature measurement, ultrasonic imaging, leakage power frequency current, and direct current components to quickly detect internal moisture and insulation defects of zinc oxide surge arresters on 10kV distribution network overhead lines, the article proposes a method and principle based on K-Means intelligent identification of defect types. Artificial samples of defects such as internal moisture and valve plate cracks of 10kV zinc oxide surge arresters are made, and a step-up device without a partial discharge source is used, Pressurize to 10kV rated voltage, collect partial discharge data through high-frequency current sensors (HFCT) and high-speed acquisition cards, and establish a database of zinc oxide defect types; By measuring 13 sets of data under different metastable states of 10kV zinc oxide lightning arresters at different measuring points on site, the results showed that the accuracy of this method in identifying internal insulation defects and internal moisture defects of 10kV lightning arresters reached over 98%, verifying the correctness of the K-Means based high-frequency partial discharge 10kV zinc oxide lightning arrester fast live detection method, which has high economic effects and social use value.