The Universal Dependencies project is currently comprised of 71 languages and 122 treebanks, and aims to find morphological and syntactic characteristics that can be applied to multiple languages for parallel language processing. In this paper, we introduce Universal POS, which is a morphological tagset for UD, and propose a method to automatically convert existing Korean morphological tagset into UPOS. In order to apply the UPOS tagset, which is based on refraction words such as English, to the Korean language, it is necessary to try a one-to-many mapping between the UPOS individual tag and the 21st century Sejong tag combination. (Yonsei University)