Osmotic dehydration of chestnut (Castanea sativa Mill.) was carried out using aqueous solutions of sucrose of different concentrations (40.0%, 50.0% and 60.0% w/w) as osmotic media, and working at three temperatures (25, 35 and 45 degrees C). At different time intervals (up to 8 h) of contact with the osmotic media, total mass, solids gain, water loss and moisture content were determined for each sample. The experimental data modelling was successfully achieved employing an empirical model based on logistic dose-response (LDR) curves considering constant dimensions. A model based on Fickian diffusion was also employed with satisfactory results considering the effective coefficient of diffusion constant. Additionally, at the same operating conditions, sucrose and moisture contents were obtained for different depths of the samples to determine composition profiles during osmotic dehydration. All the results obtained allow concluding that the osmotic dehydration kinetics depends strongly on sucrose concentration and temperature of osmotic media. At higher values for both variables, the water loss and solid acquisition are more intensive. If the water loss/ solid gain ratio as a quality parameter of the osmotic operation is analysed, working at the lowest temperature and the highest concentration assayed the best results (values around 3 after 8 h of operation) are obtained. The results on the penetration of osmotic agent indicated that the sucrose concentration and the water removal is only significant in the external slabs (4 mm. depth) at 25 degrees C being appreciable concentration gradients present in the product after osmotic dehydration. At higher temperatures, the influence of the osmotic agent is deeper. (c) 2005 Elsevier Ltd. All rights reserved.