In this paper, we propose a novel deep learning-based method for face recognition. This approach utilizes Convolutional Neural Networks (CNN) to extract facial features and a Siamese Network model to match faces. Experimental results on the LFW (Labeled Faces in the Wild) dataset demonstrate that our method achieves an accuracy of 99.2%, surpassing current state-of-the-art methods. We will discuss the applications, future directions, and limitations of this method.