This paper presents an iterative approach to solving the optimization problem of the Model Predictive Control (MPC) with constraints. The constrained cost function is formulated with the dual method and the active constraints are found at each sampling time. The minimization of the cost function is achieved by iteratively adjusting the Lagrange multipliers and the optimal control sequence using the Gradient Descent method, which reduces the computational burden because no matrix needs to be inverted. Finally, the performance of the proposed approach is tested through simulations on a multivariable 4 coupled-tank system and a comparison with the Matlab function Quadprog and the Hildreth method is made. The results show a considerable reduction in the computation time by the proposed method.