Selective oxidation of hydroxymethylfurfural to diformylfuran was performed in 3D-printed milli-scale porous reactors using pure oxygen in mild operating conditions (T = 60° C, P = 1 atm) and the homogeneous TEMPO/TBN catalytic system. Three different configurations were tested, where a rotation (θ = 22.5°) and/or an inclination (φ = 45°) of the fibers are introduced. An empty tube and a packed bed were also tested as a reference. Out of these designs, the reactor with both parameters varied simultaneously (INSP1) exhibited the highest performance, achieving an efficiency of up to 80%. The maximum conversion of 18.2% was attained for a residence time of 160 s, despite existing mass transfer limitations for this flow rate. The selectivity to DFF was 100% for all the 3D-printed reactors. On the contrary, the packed bed resulted in the highest efficiency, but at the expense of selectivity. Additional oxidation products have been retained in the packing, blocking thus the packed bed after a few hours of operation. The kinetic constant was found based on a (0,1)-order kinetic model from batch experiments. The kinetic information was utilized to evaluate the performance of the 3D-printed porous reactors from a mass transfer and reaction engineering aspect. The 3D-printed reactors were operating almost in kinetic control for total flow rates above 1 mL/min (Ha < 0.3). However, the associated short residence time resulted in small conversion. The 3D-printed reactors show significant potential when operating at higher flow rates. The low conversions can be countered by increasing the residence time, either with multiple passes or by operating them in series.