This study focuses on classifying apple varieties using machine learning techniques, leveraging the advantages of computer vision for non-destructive fruit and vegetable recognition. The research primarily uses transfer learning with seven popular CNN architectures (VGG16, VGG19, InceptionV3, MobileNet, Xception, ResNet150V2, and DenseNet201) to extract deep features. Among these, DenseNet201 achieved the highest classification accuracy of 97.48%. Deep features extracted from DenseNet201 were then used to train traditional machine learning models: support vector machine (SVM), multi-layer perceptron (MLP), random forest classifier (RFC), and K-nearest neighbor (KNN). SVM achieved the best performance with an accuracy of 98.28%, while MLP, after applying principal component analysis (PCA) to reduce dimensionality, outperformed all models with an accuracy of 99.77%. The study concludes that integrating deep features, PCA, and machine learning models can effectively classify apple varieties, with further research needed to expand the scope and usability of this technique to more varieties and larger datasets.This study focuses on classifying apple varieties using machine learning techniques, leveraging the advantages of computer vision for non-destructive fruit and vegetable recognition. The research primarily uses transfer learning with seven popular CNN architectures (VGG16, VGG19, InceptionV3, MobileNet, Xception, ResNet150V2, and DenseNet201) to extract deep features. Among these, DenseNet201 achieved the highest classification accuracy of 97.48%. Deep features extracted from DenseNet201 were then used to train traditional machine learning models: support vector machine (SVM), multi-layer perceptron (MLP), random forest classifier (RFC), and K-nearest neighbor (KNN). SVM achieved the best performance with an accuracy of 98.28%, while MLP, after applying principal component analysis (PCA) to reduce dimensionality, outperformed all models with an accuracy of 99.77%. The study concludes that integrating deep features, PCA, and machine learning models can effectively classify apple varieties, with further research needed to expand the scope and usability of this technique to more varieties and larger datasets.