The modelling of the in-duct desulfurization process at low temperatures was performed from data obtained in an entrained flow reactor at laboratory scale. Two modelling approaches were used for the process description under in-duct conditions in the reactor: (i) the application of the grain model to a single particle along the reactor, and (ii) a non-ideal adsorption equation for the gas-solid kinetics in order to solve the mass balances at a macroscopic level. It was used a synthetic flue gas stream with the typical composition of the coal-power plants, and the solid sorbent was commercial Ca(OH)(2). The experimental planning included the analysis of variables: calcium to sulfur molar ratio, SO2 inlet concentration, temperature, relative humidity, residence time, and CO2 presence in the gas stream. The experimental facility provided the measurement of SO2 concentration in the gas phase along the reactor up to 4 s of residence time for gas and solid phases. The sensitivity analysis of variables was performed by the use of a neural network (MLP 6:4:1) where the target was the experimental SO2 concentration along the reactor. (c) 2004 Elsevier B.V. All rights reserved.