This paper describes the study for the design of a predictive controller based on the parametric model obtained for a dryer cocoa plant. The drying system is a nonlinear process, the parametric model obtained is nonlinear, these conditions allowed test the performance of algorithm NEPSAC (extended, predictive and self-adaptive nonlinear controller). It shows characteristics of the plant and a brief description of nonlinear model of the plant is made. The controller design details are shown and furthermore NEPSAC results are compared with Generalized Model Predictive Control (GPC) and PID control results. The implementation of the controllers was done using MATLAB® Software and an embedded system for the collect of the data.