The main objective of this project is to study, design, develop and validate a modeling methodology for complex industrial processes that allows, on the one hand, to analyze and identify the complex dynamics of said processes, and on the other hand to offer an efficient, robust and that reproduces the behavior of the process to be controlled. The project will design, based on different structures of artificial neural networks, on the one hand, an algorithmic tool for analyzing the dynamics of nonlinear systems to predict their behavior along their equilibrium points, and on the other hand, a neural structure model that guarantees reliable and robust behavior within the scope of an Intelligent Predictive Control strategy. As a technological challenge, the project considers the implementation of the architecture of the aforementioned strategy on Hardware in the Loop (HiL) platforms, using an industrial PC and a PAC. The proposal will be validated on these HiL platforms, with a real application related to the metallurgical sector (control of a rolling process).