The AI & ML Based Pet Feeding System is a smart solution designed to identify different types of pets using image processing and deep learning techniques, specifically Convolutional Neural Networks (CNN). The system is trained on a dataset consisting of 74 different pet types using thousands of labelled images. Once trained, the model can accurately classify pets from new images uploaded by users. The application includes several modules such as dataset upload, pre-processing, train-test splitting, model training, and pet classification. After identifying the pet, the system provides feeding recommendations including suitable foods and foods to avoid, helping pet owners maintain their pet's health. The CNN model achieves high prediction accuracy, with overall performance reaching up to 93%. Graphical outputs like training accuracy, loss curves, and precision-recall for each class make the model performance transparent. This system offers an innovative and practical tool for pet identification and health-conscious feeding, making it useful for pet lovers, caretakers, and veterinary applications.