Neutrophils are directly responsible for destroying invading pathogens via reactive oxygen species, antimicrobial peptides, and neutrophil serine proteases (NSPs). Imbalance between NSP activity and endogenous protease inhibitors is associated with chronic inflammatory disorders, and engineered inhibitors of NSPs are a potential therapeutic pathway. In this study we characterized the extended substrate specificity (P4-P1) of the NSP cathepsin G using a peptide substrate library. Substituting preferred cathepsin G substrate sequences into sunflower trypsin inhibitor-1 (SFTI-1) produced a potent cathepsin G inhibitor (Ki = 0.89 nM). Cathepsin G's P2' preference was determined by screening against a P2' diverse SFTI-based library, and the most preferred residue at P2' was combined in SFTI-1 with a preferred substrate sequence (P4-P2) and a nonproteinogenic P1 residue (4-guanidyl-l-phenylalanine) to produce a potent (Ki = 1.6 nM) and the most selective (≥360-fold) engineered cathepsin G inhibitor reported to date. This compound is a promising lead for further development of cathepsin G inhibitors targeting chronic inflammatory disorders.