This paper proposes a Nonlinear Model based Predictive Control (NMPC), with guaranteed stability using a dual mode Model based Predictive Control approach with a PI controller inside a terminal region. Within the formulation of this control strategy, a terminal region (Ω) and an attraction domain are calculated using invariant sets theory and a randomized Montecarlo type algorithm. In addition, this proposal is complemented with a commutation strategy to constrain final control elements smoothing controllers commutation. This Dual Mode NMPC multivariable control is implemented by simulation over a Continuous Stirred Reactor Tank and comparing Dual Mode NMPC with a conventional NMPC multivariable and with two PI controllers. Finally, this article concludes that the NMPC for dual mode is the control strategy that in addition to having stability guaranteed, presents a better performance.