The discovery of a novel yeast with a natural capacity to produce ethanol from lignocellulosic substrates (second-generation ethanol) is of great significance for bioethanol technology. While there are some yeast strains capable of assimilating cellobiose in aerobic laboratory conditions, the predominant sugar in the treatment of lignocellulosic material, little is known about this ability in real industrial conditions. Fermentations designed to simulate industrial conditions were conducted in synthetic medium with glucose, sucrose, cellobiose and hydrolyzed pre-treated cane bagasse as a different carbon source, with the aim of further characterizing the fermentation capacity of a promising Dekkera bruxellensis yeast strain, isolated from the bioethanol process in Brazil. As a result, it was found (for the first time in oxygen-limiting conditions) that the strain Dekkera bruxellensis GDB 248 could produce ethanol from cellobiose. Moreover, it was corroborated that the cellobiase activity characterizes the enzyme candidate in semi-purified extracts (β-glucosidase). In addition, it was demonstrated that GDB 248 strain had the capacity to produce a higher acetic acid concentration than ethanol and glycerol, which confirms the absence of the Custer effect with this strain in oxygen-limiting conditions. Moreover, it is also being suggested that D. bruxellensis could benefit Saccharomyces cerevisiae and outcompete it in the industrial environment. In this way, it was confirmed that D. bruxellensis GDB 248 has the potential to produce ethanol from cellobiose, and is a promising strain for the fermentation of lignocellulosic substrates.