Freezing is one of the most employed procedures in food manufacturing and preservation, determining the quality and safety of the final frozen product, as well as the performance in further operations, often controlled by mean size and distribution of the crystals formed during the process. In this paper, two spatially distributed models based on first-principles are presented to provide a description of the freezing dynamics and the related ice crystal formation. These multi-scale models are intended to help in the understanding of the phenomena and also in the design of targeted product microstructure. A system consisting of a sucrose solution at a range of concentrations (20-60% by weight) and cooling conditions has been chosen as case study in order to demonstrate the advantages of this modelling-based approach.