Automatic traffic regulation plays an important role in public transportation especially in metro lines. Delay recovery and real-time control strategies are employed to obtain optimal schedule of high frequency lines in recent years. This paper presents a new traffic model for metro loop lines based on the time deviations of departure times. In this model, a phenomena of transferring knock on delay from one train to another one has been considered. The objective function is considered to minimize the time deviation and increase the passenger satisfaction with headway adherence. Nonlinear model predictive controller is employed to compensate the disturbances and recover the nominal time schedule in metro traffic system by varying the running time between two successive platforms. We minimize the objective function by designing nonlinear model predictive controller in the presence of the operational constraints on control actions and time interval between two successive trains. Simulation results verify the introduced model and show the performance of nonlinear model predictive controller to minimize the delays.