Immunotherapy has marked a revolution in cancer therapy. The most extensively studied target in this field is represented by the protein–protein interaction between PD-1 and its ligand, PD-L1. The promising results obtained with the clinical use of monoclonal antibodies (mAbs) directed against both PD-1 and PD-L1 have prompted the search for small-molecule binders capable of disrupting the protein–protein contact and overcoming the limitations presented by mAbs. The disclosure of the first X-ray complexes of PD-L1 with BMS ligands showed the protein in dimeric form, with the ligand in a symmetrical hydrophobic tunnel. These findings paved the way for the discovery of new ligands. To this end, and to understand the binding mechanism of small molecules to PD-L1 along with the dimerization process, many structure-based computational studies have been applied. In the present review, we examined the most relevant articles presenting computational analyses aimed at elucidating the binding mechanism of PD-L1 with PD-1 and small molecule ligands. Additionally, virtual screening studies that identified validated PD-L1 ligands were included. The relevance of the reported studies highlights the increasingly prominent role that these techniques can play in chemical biology and drug discovery.