Xylitol, a polyalcohol with anticariogenic properties, finds broad applications in different sectors. Research into new production methods has highlighted the biochemical route using immobilized microorganisms. However, challenges remain in optimizing operating conditions and understanding mass transport and biochemical reactions. Different macroscopic models have been proposed to address these challenges. Nevertheless, those models do not consider porous particles' microstructure and biofilms' formation within them, which can determine the macroscopic performance of the process due to its hierarchical nature. In this work, we derive two macroscopic models for the mass transport and reaction of the xylitol production process with immobilized microorganisms in porous particles. Such models are derived from microscopic ones using the volume averaging method, resulting in both two-equation and one-equation models, written in terms of effective medium coefficients. These latter are predicted by solving ancillary problems in representative 2D unit cells of the immobilization particles, incorporating their microstructural information. Besides, kinetic parameters are estimated through kinetic fitting using experimental data from the literature. Models' accuracy is assessed by comparing them with pore-scale simulations and experimental observations of xylitol production from sugarcane bagasse at the laboratory scale, finding good agreement. Finally, our results are compared with a macroscopic model reported in the literature, and similar predictions are found. However, unlike the reported model, the one derived here improves the modeling of the process since the effective coefficients do not need to be calculated using empirical correlations or estimators.