The Forward Modeling Controller (FMC), a recently developed model-based predictive digital controller for single-input, single-output processes, is extended to multiinput, multioutput processes. The multivariable FMC is a promising approach to the control of complex industrial processes with many inputs and outputs. The theory presented in this paper includes stability analysis plus other features necessary for robustness in industrial control. The controller has only two types of adjustments: a robustness/performance setting for each controller variable and the controller sample interval. The performance of the multivariable FMC is demonstrated on a distillation column simulation and is compared with the performance of the Dynamic Matrix Controller.