Smart nanocarriers for the transport of drugs to tumor cells are nowadays of great interest for treating cancer. The use of enzymatic stimuli to cleave peptide-based drug nanocapsules for the selective release of nanocapsules cargo in close proximity to tumor cells opens new possibilities in cancer research. In the present work, we demonstrate a methodology for finding and optimizing cleavable substrate sequences by the type II transmembrane serine protease hepsin, which is highly overexpressed in androgen-independent prostate cancer. The design and screening of combinatorial libraries in silico against the binding cavity of hepsin allows the identification of a panel of promising substrates with high-calculated docking scores. In vitro screening verifies the predictions and showed that all substrates are cleaved by hepsin with higher efficiency than the literature known hepsin substrate RQLR↓VVGG. The introduction of D-amino acids on a selected peptide with the highest catalytic efficiency (kcat/Km) renders it resistant to cleavage by plasma or serum while maintaining their susceptibility to hepsin.