There are two approaches to guarantee zero steady-state error in step reference tracking and step disturbance rejection in model predictive controllers. In the first scheme, which is more prevalent, integrator is added to the model of the system. In generalized predictive control and dynamic matrix control, this structure is used. In second scheme, the desired steady-state value of control signal is used in cost function of predictive controller. Due to substantial characteristics of second scheme, in this paper it is tried to propose an analytically tuning method for these predictive controllers. Besides, many industrial processes can be effectively described with first-order plus dead time models. Hence, in this paper these models are used to provide predictions. Using these models and considering the constraints are inactive, the closed-loop transfer function is derived and a deep study on the closed-loop system and its robustness are given. Finally, the simulation results are used to evaluate the efficiency of the proposed tuning method.