The serine protease kallikrein 8 (KLK8) is emerging as an important regulator of the neural microenvironment, playing key roles in the regulation of memory1, 2, neural plasticity3, 4, and the development of Alzheimer’s Disease5-7. Despite its attractiveness as a druggable target, there are few KLK8 inhibitors available apart from inhibitory antibodies. Previously the 14 amino acid cyclic peptide sunflower trypsin inhibitor-1 (SFTI-1), has been used as a starting point to design variants to block the activity of a range of serine proteases8-11. Here, we have used SFTI-1 as a scaffold for the development of a series of KLK8 inhibitors, using a combination of in silico molecular dynamics simulations and in vitro kinetics assays to guide substitutions within SFTI-1. Residues at positions P2’ and P5’ were optimised for KLK8 inhibition and selectivity through iterative SFTI diversity libraries of 18 natural amino acids, identifying P2’ Ile – Leu and P5’ Ile – Arg as optimal residues respectively. Incorporation of P2’ Ile – Leu into a P4-TCLR-P1 SFTI optimised for KLK8 binding12 resulted in a 3-fold increase in inhibitory potency from 10.6 nM to 3.1 nM Ki. In contrast, the incorporation of P5’ Ile – Arg resulted in substrate cleavage of the inhibitor and a dramatic decrease of inhibitory potency to 28.0 nM Ki. Though P5’ Arg can result in inhibitor-substrate conversion, molecular dynamics simulations identified similar increases in ranked binding energy, derived from the formation of a stable enzyme-inhibitor hydrogen bond network between P5’ Arg and Glu 97 of KLK8. This approach has allowed for the development of a low nanomolar inhibitor of KLK8 and provides a structural basis for future inhibitor design.