The simulated-moving-bed process is a powerful tool for continuous separation of multi-component mixtures in which the components have different adsorption affinities. It is suitable for a broad range of preparative or production scale applications and allows the separation of components with separation factors near unity with high resolution, yield and purity. Furthermore, the desorbent rates and the amount of adsorbent required by simulated-moving-bed processes are much lower than those for corresponding batch processes. The SMB process is subject to different parameters specifying plant size, operating conditions and process variables. Precise simulations are therefore necessary for process design and evaluation of experimental results. In this paper a detailed comparison of different modelling approaches is presented. The aim of this project is to develop a simulation system on the basis of a flowsheeting simulator like SPEEDUP(TM) (AspenTech, Cambridge, U.S.A.). In order to design and optimize either batch or SMB chromatographic processes it is not only necessary to describe the exact position but also the exact shape of the transient concentration fronts as a function of fluid velocity and complex component concentrations. In order to develop a simulation system which meets these demands and which is accurate enough for high-purity chromatography, appropriate models for unit operations have to be selected. Different rigorous models for batch chromatographic separations and continuous chromatography processes in terms of either moving beds with steady-state countercurrent flow or simulated moving beds with periodic fluid port switching have been evaluated. The results demonstrate that because of the real periodic flow SMB processes should be described by rigorous dynamic models with tubular reactor units and fluid port switching. Axial dispersion and mass transfer resistance have to be taken into account in order to obtain a good agreement with experimental results. The rigorous dynamic model which has been developed makes it possible to understand malfunctions of pilot plants, to optimize process conditions and to specify startup procedures for SMB processes. (C) 1998 Elsevier Science Ltd. All rights reserved.