A previously developed pseudo-component adaptive characterization method was applied to the multistage separation of continuous mixtures. The objective of this new approach is to provide an adaptive scheme for the process simulation of continuous multicomponent mixtures that can be accurate using few pseudo-components. For illustration purposes, the separation of continuous and discrete mixtures of hydrocarbons in distillation columns was simulated using a sequential flash approach. Phase equilibrium was calculated either by Raoult's law or by using cubic equations of state. The simulation results were compared to those obtained by the traditional bubble-point method in which the mixture characterization in pseudo-components was performed by the fixed-point Gaussian quadrature method. The new method showed a remarkable accuracy, giving results similar to those obtained by the traditional method but with an almost order of magnitude reduction in the number of pseudo-components. Therefore, this adaptive characterization method may potentially be applied as an order reduction method to computationally intensive applications involving continuous mixtures, such as CFD simulations and optimization problems. (c) 2012 Elsevier B.V. All rights reserved.