A flexible and comprehensive model for predicting and optimizing separations of racemates of amino acids and other chiral metabolites in liquid-liquid multi-staged extraction systems is presented. Enantiomer partition coefficients are computed along the extraction path using multiple chemical equilibria theory and measured equilibrium formation constants for every complex formed in the two phases. The large number of speciation reactions typically occurring in ligand-exchange extraction systems requires the development of a robust numerical algorithm, and we present a method to rapidly and accurately solve the large nonlinear set of governing equations. Model performance is assessed through comparison to data for continuous extraction of various racemates within a series of hollow-fiber membrane modules. For each extraction, a chiral-ligand exchange selector molecule is solubilized in the organic phase flowing countercurrent to the aqueous phase into which the racemate is loaded. Enantiomer eluent profiles predicted at different conditions are in very good agreement with experiment. Through its predictive power, the model provides a useful in silicoplatform for optimizing these complex separations, and model results demonstrating this capability are presented.