One critical point for the successful installation of an industrial MPC control system is its ability to operate acceptably under different operational scenarios that occur over time in all process industries. e, mainly because they lack the required robustness to operate under different scenarios that may appear in industrial practice. In the core of many MPCs there is a fix multivariable linear model that has been obtained for specific operating conditions, so that its performance changes when its operation moves to a different point. To address this well-known problem, several solutions have been proposed in the literature. In this communication, we present a method to modify the internal model of these multivariable controllers so that their behavior do not present abnormalities that force the operators to set them off, when a change in the operating mode takes place. The approach modifies the model taking into account the directional gains of the operating scenarios while keeping it as close as possible to the original identified one. The communication describes the method proposed, named Gain Consistency, and illustrates its application in a typical process unit: a distillation tower of an oil refinery.