This work deals with the creation of a mathematical model which will predict with sufficient accuracy the size of granules in the high-shear granulation process. The aim of the paper was to create a model created with dimensionless criteria to give the size of granules as a function of selected process parameters, equipment, and experimental material. Microcrystalline cellulose granules were manufactured in an experimental granulator of our own construction using a high amount of granulation liquid (absolute humidity >40%). The work also focuses on the study of granule size distribution and the yield of the process in achieving the target granule fraction (0.4-3 mm). Particle size and particle size measurement were performed using 3 D dynamic image analysis (PartAn 3 D, Microtrac). The granules were analyzed with respect to their shape and width of distribution. Using the created model, it is possible with high probability, 89.4%, to predict on our device the median value of the granule size of microcrystalline cellulose, with +/- 0.1 mm precision.