Torrefaction is a promising technology to produce woody sustainable materials and biochar for combustion and gasification. To aid in the process and reactor design concept, numerical models are applied to predict processes parameters giving treatment time estimation, solid and volatile yields, and calorific values of the solid fuels. In this work, a hardwood (Eucalyptus grandis) was investigated by using thermogravimetric analysis (TGA), as well as elemental analysis, which can reveal the thermal events and identify degradation products as a function of treatment time and temperature. A numerical prediction based on a two-step kinetics model was employed to predict solid yield, elemental composition, higher heating values (HHV), and its enhancement factor during the treatment. The further analyzed prediction 3D surfaces and profiles allows clear identification of the torrefaction severity. The predicted HHV's showed enhancement factors up to 10.2% for severe torrefaction. The obtained kinetic rates for Eucalyptus grandis, as well as the numerical results, can provide useful information for industrial operation and reactor projects.