There are many types of Korean characters, and it is difficult to effectively extract highly discriminative features when processing continuous text or long strings, resulting in low accuracy in Korean character recognition. Therefore, in order to optimize the intelligent recognition effect of Korean characters and improve the accuracy of character recognition, a new Korean character intelligent recognition algorithm was studied by using deep learning to improve long short-term memory. Firstly, preprocess single characters in Korean to transform character images into a unified form; Secondly, useful features, including texture and shape features, are extracted from the preprocessed character images to more comprehensively describe the characteristics of Korean characters; Using deep learning to improve long short-term memory and capture the sequence relationships of Korean characters, in order to improve the coherence and accuracy of recognition when processing continuous text or long strings; On this basis, calculate the similarity of character vectors and intelligently classify and recognize Korean characters. The test results show that the proposed algorithm performs the best in both character recognition accuracy and line recognition accuracy, both reaching over 98%, with significant advantages in recognition performance.