We report the combination of measurement and prediction of X-ray absorption fine structure (XAFS) data, where the term XAFS refers to the overall spectrum that encompasses both the X-ray Absorption Near Edge Structure (XANES) region as well as the Extended X-ray Absorption Fine Structure (EXAFS) region, to evaluate the cadmium thiolate cluster structures in the metalloprotein metallothionein. XAFS spectra were simulated using coordinates from molecular models of the protein calculated by molecular mechanics/molecular dynamics (MM3/MD), from NMR analyses, and from analysis of X-ray diffraction data. XAFS spectra were also simulated using the coordinates from X-ray crystallographic data for [Cd(SPh)(4)](2-), CdS, [Cd-2(mu-SPh)(2)(SPh)(4)](2-), and [Cd-4(mu-SPh)(6)(SPh)(4)](2-). The simulated XAFS data that were calculated using the FEFF8 program closely resemble the experimental data reported for [Cd(SPh)(4)](2-), CdS, [Cd-2(mu-Sph)(2)(SPh)(4)](2-), [Cd-4(mu-SPh)(6)(SPh)(4)](2-), rabbit liver metallothionein cadmium alpha-domain (Cd-4-alpha MT), and cadmium rabbit liver beta alpha metallothionein (Cd-7-beta a MT). MM3 force field parameters were modified to include cadmium-sulfur bonding and were initially set to values derived from published X-ray diffraction and EXAFS experimental data. The force field was further calibrated and adjusted through comparison between experimental spectra taken from the literature and simulated XAFS spectra calculated using the FEFF8 program in combination with atomic coordinates from MM3/MD energy minimization. MM3/MD techniques were used with the calibrated force field to predict the high-resolution structure of the metal clusters in rabbit liver Cd-7-MT. Structures for Cd3S9 (beta) MT and Cd4S11 (alpha) MT domains from MM3/MD calculations and those previously reported for Cd7MT on the basis of H-1 and Cd-113 NMR data were compared. Structural differences between the different models for these cadmium thiolate clusters were evident. Combining the measurement and simulation of XAFS data provides an excellent method of assessing, modeling, and predicting metal-binding sites in metalloproteins when X-ray absorption spectroscopy (XAS) data are available.