During electrorefining of spent nuclear fuels, fission products, and actinides accumulate in the LiCl-KCl electrolyte salt. From the standpoint of high-level waste minimization, it is advantageous to remove these from the salt rather that discarding the salt after these have built up to certain concentrations. Laboratory experiments have shown that zeolite A has the desirable properties for selective removal of fission products by the ion-exchange process. An analytical model has been developed to assess the performance of the ion-exchange column and to scale up the data to engineering size equipment. The model employs empirically determined exchange factors to represent the equilibrium chemistry of the exchange process. The exchange kinetics is modeled by the Nernst-Planck equation applied to multicomponent diffusion in the zeolite pellet. The self-diffusion coefficients, derived from batch kinetic data, are substantially different for monovalent, divalent, and trivalent fission products. The method of characteristics is used to solve the transport equations for the salt phase. The computed breakthrough of the fission products through the experimental zeolite column is in partial hut not complete agreement with the measured concentration profiles.