An array of 4000 defined and addressable tripeptides on a polymer-coated glass slide is used to synthesize molecularly imprinted polymer (MIP) nanoparticles. This work is undertaken to systematically probe the impact of the peptide sequence on the ability to generate affinity MIPs. The polymer affinity is assessed by measuring the fluorescence of bound MIP nanoparticles at each peptide spot on the surface after washing the array to remove any low-affinity polymer. The generic composition commonly used in the preparation of MIPs against proteins seems to be equally suitable for imprinting hydrophobic and hydrophilic tripeptides. The amino acids frequently contributing to the formation of high-affinity MIPs include T, F, D, N, Y, W, and P. The amino acids that rarely contribute to the formation of high-affinity interactions with MIPs are G, V, A, L, I, and M. These observations are confirmed by computational modeling. The basic technique proposed here may be applicable in optimizing polymer compositions for the production of high-affinity MIPs or, more specifically, for the selection of appropriate amino acid sequences when peptide epitopes are used instead of whole protein imprinting.