The solid oxide fuel cell power plant is known to be a potential alternative in the electric utility. However, the output voltage of the solid oxide fuel cell changes with the load variations. Model predictive control is part of a family of optimization-based control methods, which are based on online optimization of future control moves. This paper proposes a model-based controller for the regulation of a solid oxide fuel cell. The performances using both the linear and fuzzy Hammerstein models are evaluated with constraints