As part of safety monitoring, checking the quality of peanuts is indispensable. Peanuts contain oils and proteins that are denatured at elevated moisture levels and undergo vigorous metabolism, resulting in mass decay during storage. In this study, the e-nose signals of peanut kernel were classified into three categories of high quality, medium quality and low quality for each storage period by principal component analysis (PCA), which was consistent with the physical and chemical quality clustering results. Using the classification results as learning labels, support vector machine (SVM), decision tree, linear discriminant analysis (LDA) and k-nearest neighbor learning (KNN) were used to develop quality discrimination models for peanut kernels, and the prediction accuracy of each model could reach more than 90%. The results show that the e-nose has an excellent ability to discriminate the quality changes of peanut kernels during storage, in agreement with its classification of physical and chemical quality changes.