A three-step bio-oil production process involving torrefaction pretreatment (225, 250, or 275 °C at 20 min hold time), pyrolysis (500 °C, two heating rates), and secondary catalytic processing over HZSM-5 (at 400, 450, or 500 °C) was studied to determine how treatments affected byproduct formation and catalyst properties and function. Torrefaction pretreatment significantly reduced average yields (%, w/w of feed) of reactor char (100% reduction), catalyst coke (21.4%), and catalyst tar (8.1%) relative to the best-case conditions using non-torrefied feedstock. The yields of chemical components, including levoglucosan, formic acid, acetic acid, and 5-hydroxymethylfurfural (5-HMF), were significantly reduced in intermediate fast-pyrolysis bio-oil (FPO) derived from torrefied feedstock, while for intermediate slow-pyrolysis bio-oil (SPO), yields of levoglucosan, acetic acid, and 5-HMF were reduced. However, in terms of concentration, only furfural showed a significant correlation with the torrefaction temperature. Both furfural and formic acid indicated correlations with coke formation, although these were negative correlations. For formic acid, a decreased coke yield with an increasing concentration was attributed to the formation of H 2 from the thermal decomposition. Combined coke, char, and tar yield significantly decreased with an increasing torrefaction temperature, decreasing formic acid, acetic acid, and furfural concentrations, and an increasing levoglucosan concentration to a minimum of 14.4% (w/w of bio-oil feed). Torrefaction also increased catalyst effectiveness for minimizing changes to pore size, pore volume, and surface area upon upgrading FPO but reduced effectiveness for SPO processing. Torrefaction of biomass prior to slow pyrolysis more effectively maintained weak and strong acid site density after catalytic processing, although no clear effect of torrefaction was seen for FPO-processing catalysts.