Through coordinate transformation, the predictive model of the inverter output currents is proposed in the two-phase static αβ-frame. The cost function is used to assess different voltage vectors of the inverter and predict all possible output currents. The voltage vector which minimizes the cost function is selected. Then the corresponding switching state is applied at the next sampling time to track reference currents quickly. The proposed strategy needn't to use the traditional PWM modulator and the inner current loop. Only once coordinate transformation is needed to calculate the optimal switching state, so the strategy has small amount of calculation and easy to be implemented. Simulation results explain that if reference currents change dynamicly, model predictive current control can track reference currents quickly. If the light intensity changes, the model predictive current control can track the maximum power point rapidly and the inverter output currents have low harmonic distortion. The simulation results verify the effectiveness of the proposed control strategy by comparing the output waveforms under SVPWM control.