This paper presents a nonlinear predictive controller based on state space model applied to a dynamic model with partial equations of refrigeration for mango preservation. Refrigeration systems are widely used in the food industry and are essentially important for the Perú's agro-export sector, the main products in this sector are mango, banana, and grape. Advanced control strategies offer the potential to increase the energy efficiency of refrigeration processes while ensuring the quality required for agro-export products. This paper proposes a NMPC designed from a dynamic model based on the fundamentals of thermodynamics and heat transfer of all the elements involved in this system, implemented in a simulated plant from a mathematical model with a switching mode of operation in the evaporator that was previously validated to ensure a similar behavior to a real plant and has been compared with a PID controller to check the control efficiency.