Abstract—Nowadays, although Model Predictive Control has been interested in stabilizing systems, it mainly considered terminal equality constraints method by adding a terminal – state penalty into the cost function (i.e finite predictive horizon). But up to now there is no any law to find a terminal – state penalty. The Twin Rotor MIMO systems (TRMS) is a nonlinear object with a complex dynamic. This paper offers a method which differs one based on the Bellman’s dynamic programming for stably tracking of bilinear continuous systems with infinite predictive horizon instead of controlling stably and uses continuous model directly instead of non-continuous approximate model. This method is built basing on variation of optimal control and apply to control TRMS.