In this paper, a nonlinear receding-horizon estimation scheme is proposed for general nonlinear diffusive binary chromatographic processes. This is done by approximating the. concentration profiles over a suitable truncated functional basis. The resulting vector of unknowns is first reduced using available measurements. Then the remaining part is computed by nonlinear optimization. Numerical simulations are proposed to illustrate the efficiency of the proposed scheme. (C) 2003 Elsevier Science Ltd. All rights reserved.