Accurately predicting the internal hot spot temperature can improve the damp fault inspection effect of arrester effectively. An inversion detection method of hot spot temperature inside 500 kV arrester based on surface temperature and SVM (support vector machine) is proposed. Taking the surface temperature of arrester and wind speed as inputs, the inversion of hot spot temperature inside arrester is realized. In order to improve the prediction accuracy of the proposed model, the influences of the grid search (GS) and particle swarm optimization (PSO) algorithms on inversion accuracy are compared. The results show that the GS?SVM model is of good inversion performance. The maximum and minimum errors between the internal hot spot temperature obtained by the inversion method and the actual values are 4.00 ℃ and 0.01 ℃ respectively, which proves the effectiveness of the inversion model.