A dynamic model describing the transfer of mass, energy and momentum has been developed for a rotary dryer. The model is applied to an industrial case, drying of sugar beet shreds at a Danish sugar factory. The model is fitted to plant data and shows a good ability in predicting variations in product quality with changes in manipulated variables and disturbances. A comparison is made, in a simulation study, with traditional feedback control and model based predictive control (MPC) with feedforward action.