A selective fractionation method of carbohydrate mixtures of galactose/tagatose, using supercritical CO(2) and isopropanol as cosolvent, has been evaluated. Optimization was carried out using a central composite face design and considering as factors the extraction pressure (from 100 to 300 bar), the extraction temperature (from 60 to 100 degrees C), and the modifier flow rate (from 0.2 to 0.4 mL/min, which corresponded to a total cosolvent percentage ranging from 4 to 18% vol). The responses evaluated were the amount (milligrams) of tagatose and galactose extracted and their recoveries (percent). The statistical analysis of the results provided mathematical models for each response variable. The corresponding parameters were estimated by multiple linear regression, and high determination coefficients (>0.96) were obtained. The optimum conditions of the extraction process to get the maximum recovery of tagatose (37%) were 300 bar, 60 degrees C, and 0.4 mL/min of cosolvent. The predicted value was 24.37 mg of tagatose, whereas the experimental value was 26.34 mg, which is a 7% error from the predicted value. Cosolvent polarity effects on tagatose extraction from mixtures of galactose/tagatose were also studied using different alcohols and their mixtures with water. Although a remarkable increase of the amount of total carbohydrate extracted with polarity was found, selective extraction of tagatose decreased with increase of polarity of assayed cosolvents. To improve the recovery of extracted tagatose, additional experiments outside the experimental domain were carried out (300 bar, 80 degrees C, and 0.6 mL/min of isopropanol); recoveries >75% of tagatose with purity >90% were obtained.