Abstract: Adoptive cell immunotherapy with T cell receptor-engineered T cells(TCR-T) targeting immunogenic tumor antigens can effectively kill tumors without obvious side reactions. However, this field is still in its infancy, with great challenges to how to screen tumorspecific antigens and their corresponding T cell receptors(TCRs). In this paper, the advances in research on current TCR-T clinical trials are reviewed, the development of machine learning involved in putative antigen and candidate TCR prediction is summarized, and different screening strategies for tumor-specific antigens and corresponding TCRs based on comprehensive studies are introduced, so as to provoke new insight for further development of TCR-T therapy.