This work addresses the optimal design of a flexible heat exchanger network using model-based optimization, applied to hydrogen production by means of an ethanol steam reforming process. High efficiencies are obtained at different hydrogen production levels ranging from 25 to 100% of a nominal output. System structure, heat exchanger sizing, and operation conditions are simultaneously settled, ensuring both operational feasibility and optimality. The system involves a reforming reactor, vaporization and reheating equipment, combustors, and a heat exchanger network system. A multi-period nonlinear optimization problem (NLP) was formulated to account for the production level distribution. Equipment sizing constraints and structural constraints link the different scenarios. The trade-off between area and efficiency is analyzed using a multi-objective epsilon-constraint approach. Models were developed in the GAMS environment. The resulting solutions, for the maximum area case, maintain alcohol combustion at low levels showing efficiencies around 63% in each operational level. Pareto Optimal diagram shows that a 1% reduction of efficiency allows a 50% decrease in total required heat exchanger area by 50%.