Efficient conversion of food waste to value-added products necessitates the development of high-performance heterogeneous catalysts. This study evaluated the use of Al2O3 as a low-cost and abundant support material for fabricating Lewis acid catalysts, i.e., through the in-situ doping of Cu, Ni, Co, and Zr into Al2O3 followed by calcination. The characterisation results show that all catalysts were mainly amorphous. In particular, adding the transition metals to the Al2O3 matrix resulted in the increase of acidity and meso-/micro-pores. The catalysts were evaluated in the conversion of glucose, which can be easily derived from starch-rich food waste (e.g., bread waste) via hydrolysis, to fructose in biorefinery. The results indicate that the Ni-doped Al2O3 (Al-Ni-C) achieved the highest fructose yield (19 mol%) and selectivity (59 mol%) under heating at 170 °C for 20 min, of which the performance falls into the range reported in literature. In contrast, the Zr-doped Al2O3 (Al-Zr-C) presented the lowest fructose selectivity despite the highest glucose conversion, meaning that the catalyst was relatively active towards the side reactions of glucose and intermediates. The porosity and acidity, modified via metal impregnation, were deduced as the determinants of the catalytic performance. It is noteworthy that the importance of these parameters may vary in a relative sense and the limiting factor could shift from one parameter to another. Therefore, evaluating physicochemical properties as a whole, instead of the unilateral improvement of a single parameter, is encouraged to leverage each functionality for cost-effectiveness. This study provides insights into the structure-performance relationships to promote advance in catalyst design serving a sustainable food waste biorefinery.