Various applications of image processing in agriculture are extensively explored by researchers. Fruit image classification plays a key role in agricultural and food industries. Accurate identification of fruits is important for quality control, sorting, and inventory management. Automating fruit classification requires advanced techniques to handle different shapes, sizes, and colors. Deep learning has become a powerful tool for this task. It extracts complex patterns from images. Among deep learning models, AlexNet is proven to be effective in image recognition tasks. Its layered architecture efficiently captures visual features and handles large datasets, making it ideal for fruit classification problem. This research proposes an AlexNet-based deep learning model for fruit image classification. The model is designed to enhance classification accuracy by leveraging AlexNet’s strength in feature extraction. A fruit image dataset comprising 100 different fruit images with over 90 K images has been used to train the AlexNet model. Other deep models such as LeNet-5, SqueezeNet, and VGG16 have also been investigated for fruit image classification. Experimental results showed that AlexNet has achieved a recall value of 0.964. Hence, it is excellent at identifying most of the fruit images. The accuracy of 0.974 indicates that AlexNet has correctly classified almost all images. The precision value 0.940 indicates that AlexNet is also good at minimizing false positives.