This paper presents a detailed description of Finite Control Set Model Predictive Control (FCS-MPC) applied to three-level neutral point clamped inverter (3L-NPC). The controller uses a model of the converter, DC link and load to calculate predictions of the future values of the load currents and DC-link capacitor voltages for all possible voltage vectors. The cost function is defined to minimize the error between the reference currents and predicted load currents, balance the DC-link capacitor voltage and reduce the switching frequency. The optimal switching state that minimizes the cost function is selected and applied to the converter. Furthermore, the quality of the current and voltage are improved by using a two prediction steps procedure. Simulation results for the one and two prediction step are presented and compared with linear current controller and space-vector modulation (PI-SVM). The obtained results show the better performances of the proposed control method.