Mannitol is a natural low-calorie sugar alcohol produced by certain (micro)organisms applicable in foods for diabetics due to its zero glycemic index. In this work, we evaluated mannitol production and yield by the fruit origin strain Fructobacillus tropaeoli CRL 2034 using response surface methodology with central composite design (CCD) as optimization strategy. The effect of the total saccharide (glucose + fructose, 1:2) content (TSC) in the medium (75, 100, 150, 200, and 225 g/l) and stirring (S; 50, 100, 200, 300 and 350 rpm) on mannitol production and yield by this strain was evaluated by using a 2(2) full-factorial CCD with 4 axial points (alpha = 1.5) and four replications of the center point, leading to 12 random experimental runs. Fermentations were carried out at 30 A degrees C and pH 5.0 for 24 h. Minitab-15 software was used for experimental design and data analyses. The multiple response prediction analysis established 165 g/l of TSC and 200 rpm of S as optimal culture conditions to reach 85.03 g/l [95% CI (78.68, 91.39)] of mannitol and a yield of 82.02% [95% CI (71.98, 92.06)]. Finally, a validation experiment was conducted at the predicted optimum levels. The results obtained were 81.91 g/l of mannitol with a yield of 77.47% in outstanding agreement with the expected values. The mannitol 2-dehydrogenase enzyme activity was determined with 4.6-4.9 U/mg as the highest value found. To conclude, F. tropaeoli CRL 2034 produced high amounts of high-quality mannitol from fructose, being an excellent candidate for this polyol production.