The ethanol accumulated in the broth during fermentation is the main component toxic to yeast, causing slower yeast growth and decreased ethanol production. One way of overcoming this inhibition effect is to use extractive fermentation, where the ethanol is removed from the broth during the fermentation process. The present work evaluates ethanol production by extractive fed-batch fermentation with CO2 stripping, under different conditions of substrate concentration in the must feed (CsF), vat filling time (Ft), and start time of ethanol stripping with CO2. First, the process kinetic parameters were estimated by modeling of conventional fed-batch fermentations (without stripping) in a 5 L bubble column bioreactor, with fitting of the model to experimental data. This procedure used a sucrose concentration of 180 c in the must feed, temperature of 34.0 degrees C, and vat filling times of 3 and 5 h. Subsequently, extractive fed-batch ethanol fermentations were performed at 34.0 degrees C with a sucrose concentration of 180 g.L-1 in the feed, specific CO2 flow rate (phi) of 2.5 vol.vol(-1).min(-1) (vvm), and Ft of 3 or 5 h, starting ethanol stripping with CO2 after 3 or 5 h of fermentation. The hybrid Andrews-Levenspiel model was able to provide accurate descriptions of the behaviors of the conventional and extractive fed-batch ethanol fermentations, considering the removal of ethanol and water from the broth. Use of Ft of 5 h and start of ethanol stripping at 3 h of fermentation substantially reduced the inhibitory effects of the substrate and ethanol on the yeast cells. This condition enabled the extractive fed-batch ethanol fermentation to be performed using substrate concentrations of up to 240 g.L-1 in the feed, with substrate exhaustion occurring after approximately 12 h. The total ethanol concentration reached 110.3 g.L-1 (14 degrees GL (degrees Gay-Lussac)), around 33% higher than that obtained using conventional fed-batch fermentation without ethanol removal.