Distribution cables are an important component of the power system, and partial discharge caused by insulation defects is the main cause of cable failures. This article proposes a multi feature partial discharge prediction method based on Long Short Term Neural Network (LSTM) for cable terminal ground potential spikes. By measuring all partial discharge data from the start of partial discharge to insulation breakdown of cable terminal ground potential spikes, the development trends of parameters such as maximum discharge capacity, average discharge capacity, discharge power, and square rate were obtained; Developed a multi parameter partial discharge prediction analysis based on LSTM neural network, achieving active prediction of partial discharge for this defect.