Many proteins bind transition metal ions as cofactors to carry out their biological functions. Despite wild-type binding affinities for transition metal ions being predominantly dictated by the Irving- Williams series, in vivo metal ion binding specificity is ensured by intracellular mechanisms that regulate free metal ion concentrations. However, a growing area of biotechnology research considers the use of metal-binding proteins in vitro to purify specific metal ions from wastewater, where specificity is dictated by the protein's metal binding affinities. A goal of metalloprotein engineering is to modulate these affinities to improve a protein's specificity towards a particular metal; however, the quantitative relationship between the affinities and the equilibrium metal-bound protein fractions depends on the underlying binding kinetics. Here we demonstrate a high-throughput intrinsic tryptophan fluorescence quenching method to validate kinetic models in multi-metal solutions for CcNikZ-II, a metal-binding protein from Clostridium carboxidivorans. Using our validated models, we quantify the relationship between binding affinity and specificity in different classes of metal- binding models for CcNikZ-II. We further demonstrate that principles for improving specificity through changes in binding specificity are qualitatively different depending on the competing metals, highlighting the power of mechanistic models to guide metalloprotein engineering targets.