This paper presents a deep learning-based handwriting recognition system using Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), and Connectionist Temporal Classification (CTC). Unlike traditional OCR methods, which struggle with irregular handwriting, our approach combines robust preprocessing, CNN-RNN feature extraction, and advanced decoding strategies such as Beam Search and Word Beam Search. With a word accuracy of 92%, this model works well with different styles of handwriting. Uses for this technology include digitizing printed documents, automatic form-filling, and accessibility features. The experiment shows that deep learning not only addresses problems with handwriting recognition but also notes where future experiments can apply real-time implementation and larger data sets for increased efficiency in the real world.