Abstract: The spray method is a common way to detect hydrophobicity of composite insulators. However, it has some disadvantages such as low efficiency and poor accuracy since it totally depends on human judgment. To overcome this problem, an intelligent hydrophobicity classification method based on convolutional neural network (CNN) is proposed in this paper. Firstly, we simulated different hydrophobicity classes (HCs) by spraying different concentrations of ethanol solution on composite insulators. Under each HC, the hydrophobicity images of clean, fouled and discolored insulators were obtained in consideration of real conditions such as different photograph angles, different photograph distances and different light intensity. In order to reduce the computational complexity and improve the classification accuracy, the hydrophobicity image is preprocessed by cutting, compressing and enhancing, and then it is input into the CNN model designed for image feature extraction and classification, so as to realize the intelligent classification of insulator hydrophobicity. The results show that the CNN-based on composite insulator hydrophobicity intelligent classification method can effectively identify each HC under real complex conditions, the accuracy can reach more than 90%. It has good generalization ability and a certain application potential.