Vietnamese ancient literature, including Han works and Nom works, have lasted for hundreds of years, preserving a wealth of cultural heritage and invaluable knowledge. However, the complexity of Han-Nom texts makes them inaccessible to even most Vietnamese readers. In this paper, we introduce a machine translation system designed to translate Vietnamese ancient texts into English, bridging the gap between these rich historical works and a broader audience. Given the difficulty of building parallel data corpus for these texts, we propose a data augmentation technique using Large Language Models (LLMs) approach combined with back translation. This enriched dataset is then used to fine-tune the EnViT5 model developed by VietAI. Our experiments show that the model achieve a SacreBLEU score of 5.870 and a BLEURT score of 34.349.