Animal recognition and identification research is critical for animal protection and monitoring. We focused specifically on cats in this report. A method for cat recognition and identification is proposed that is based on Autoencoders and Convolutional Neural Networks (CNN). Additionally, we created a fresh dataset containing 1,994 photos of 17 cats. This article details how to utilize Autoencoder to denoise cat image data and combine it with CNN to create a strong model for the same cat recognition, which serves as motivation and a reference for future research in animal recognition.