The paper introduces HGTDRI (Heterogeneous Graph Transformer for Drug Repurposing), a novel approach to drug repurposing that addresses the limitations of existing methods. HGTDRI is a three-step process: constructing a heterogeneous knowledge graph, utilizing a heterogeneous graph transformer network, and computing relationship scores using a fully connected network. The method leverages the PrimeKG knowledge graph, which integrates 20 high-quality resources, and incorporates BioBERT and ChemBERTa embeddings to enhance the initial embeddings. The heterogeneous graph transformer (HGT) technique is used to extract node features, and a fully connected network predicts the relationship score between drugs and diseases. The evaluation demonstrates that HGTDRI performs comparably to previous methods and shows promising results in predicting drug repurposing candidates. The method is also validated through medical studies and demonstrated to predict other types of relationships, such as drug-protein and disease-protein interactions. The source code and data are available online.The paper introduces HGTDRI (Heterogeneous Graph Transformer for Drug Repurposing), a novel approach to drug repurposing that addresses the limitations of existing methods. HGTDRI is a three-step process: constructing a heterogeneous knowledge graph, utilizing a heterogeneous graph transformer network, and computing relationship scores using a fully connected network. The method leverages the PrimeKG knowledge graph, which integrates 20 high-quality resources, and incorporates BioBERT and ChemBERTa embeddings to enhance the initial embeddings. The heterogeneous graph transformer (HGT) technique is used to extract node features, and a fully connected network predicts the relationship score between drugs and diseases. The evaluation demonstrates that HGTDRI performs comparably to previous methods and shows promising results in predicting drug repurposing candidates. The method is also validated through medical studies and demonstrated to predict other types of relationships, such as drug-protein and disease-protein interactions. The source code and data are available online.