This paper presents a model and its identification, which characterizes the thermal dynamic heating cycle for No. 22 Hot Blast Stove, located in the No. 2 Blast Furnace, at CompaaSiderrgicaHuachipato in Talcahuano, Chile. System identification was used to model the dynamics for heating cycle based on the empirical data from the plant. In addition, a control system is proposed to obtain heating curves that generate energy savings. Model Predictive Control (MPC) was employed to minimize the tracking error to the reference while satisfying operation constraints. The feasibility of the proposal is verified according to the simulation results, estimating a 4% savings in fuel usage. © 2011 IEEE.