HPV16-associated cancers require effective T-cell-based therapeutic vaccines. In this study, we applied machine-learning approaches to prioritize HPV16 T-cell epitopes and identify promising candidates for future vaccine development. Experimentally validated HPV16 epitopes, their corresponding HLA data, and T-cell assay results were obtained from the Immune Epitope Database (IEDB). Based on the amino-acid sequences, physicochemical and molecular-weight-related features were derived, together with BLOSUM62 sequence representations and immunological properties such as HLA-specific binding information.  LightGBM and Random Forest models were trained to predict immunogenicity scores and rank candidate epitopes. LightGBM demonstrated superior overall performance across multiple evaluation metrics. Feature-importance analysis indicated that BLOSUM62-based sequence encoding contributed most strongly to model predictions, while HLA-related and physicochemical properties provided additional information.  The resulting model successfully prioritized HPV16 epitopes in an HLA-specific manner and identified high-scoring candidates, including epitopes with limited prior experimental evidence. These findings demonstrate the potential of machine-learning-based approaches for the systematic prioritization of HPV16 T-cell epitopes. Future work will focus on motif analysis and experimental validation of selected candidates.