The feasibility of large-scale separation of amino acids and inorganic ions by ion exchange columns has been demonstrated by considering a prototype problem involving one amino acid (proline), sodium chloride, and chloridic acid. The optimal design and operation of such a process has been based on an appropriate mathematical model, which includes the description of multicomponent ion exchange equilibria and the simulation of inter- and intraphase transport processes. The latter is based on a new lumping approximation of the Nernst-Planck equation for the diffusion of ions. The model contains some adjustable parameters that have been estimated from binary equilibrium data and experimental binary breakthrough curves of all the involved ionic species. The model reliability has been tested by comparison with experimental effluent histories of several adsorption-elution cycles involving multicomponent mixtures. Using the model, an optimal column operation, which includes ammonia feed before regeneration, has been designed for the separation of the prototype mixture. This result has been tested experimentally at the laboratory scale.