The most commonly used tuning strategy for a model predictive control MPC is by introducing a move suppression coefficient to prevent excessive changes in manipulated variables. However this approach involves the iterative selection of certain parameters that are not well defined and therefore demands sound understanding of the theoretical formulations. This procedure may lead to suitable process close loop responses. A more advanced approach is the use of constrained optimization methods which are computationally demanding in nature making it less suited for tight control of fast processes. In this paper an effective tuning strategy for predictive control i.e. shifting method is proposed which leads to the reformulation of the original predictive control structure. The inherent ill-conditioning is eliminated by allowing the process prediction time step to be decoupled from the control time step. The original open loop data is used to evaluate a shifting factor m to be applied to the dynamic matrix structure which replaces the move suppression coefficient. The results show that the proposed tuning strategy gives improved closed loop responses for control simulations on a multivariable non-linear process having variable dead-time and on other models found in the literature. The algorithm was also practically demonstrated on a fast reacting process and multi input multi output MIMO slow reacting plant i.e. DC motor rotational speed control and a pilot scale distillation column respectively with better control being realized in comparison with DMC using move suppression. A major benefit of this proposed method is that only minor modification is required in order to implement this tuning strategy into the existing un-constrained control algorithm. It also eliminates the need for computationally intensive optimization of move suppression and uses purely open loop process data for tuning. The shifting factor m is generic therefore it can be effectively applied for any control horizon and any processes.