This paper presents the dynamic modeling and control of a rotary disc dryer used in fishing industry. The nonlinear model obtained from energy and mass balances, is validated with real data measured in two plants of fishmeal located in the north of Perú. A constrained nonlinear predictive controller is designed with the purpose of regulating the moisture content in the fishmeal to obtain a stable final product. Simulation results are shown to illustrate the effectiveness of the controller to handle input, slew-rate and output constraints even for the case of short prediction horizons.