The stockbridge damper identification is of the great significance of the overhead power line inspection. To improve inspection efficiency, this paper proposes an identification method based on Gray Wolf Optimization (GWO), histogram of oriented gradient (HOG) feature and Support Vector Machine (SVM). Firstly, the HOG feature of the data set is extracted as input. GWO is used to optimize parameters c and g. Then, SVM is trained by the optimal parameters to realize the automatic identification of stockbridge damper. Besides, this paper proposes a slide window detection method based on the position of the power line for detection in large scenes. The position is determined by pre-processed methods, and the identification speed is improved. The results show that GWO-SVM can identify the stockbridge damper effectively, and furthermore this method has higher recognition accuracy than the SVM optimized by Grid Search (GS) and Particle Swarm Optimization (PSO).