In order to balance the output current of each charger related with energy storage type light rail vehicles, an explicit predictive current-sharing control is proposed in this paper. The dynamics of the parallel charging system with non-identical node is formulated as a multi-input and multi-output state space model, which is a second-order nonlinear system. As the duty ratio input of each charger is bounded and the current should evolve within the range of constraint, the current-sharing control design can be converted into a model predictive optimization problem. The cost function related with this nonlinear optimization problem includes three terms named the output current tracking error energy, the control input changing error energy and the output current cooperative interaction energy. By adopting the proposed predictive current-sharing control, the output current of each charger can track the reference current in a finite time while the duty ratio of each charger can satisfy the boundedness constraint and the current can also satisfy the peak value constraint at the same time. Simulation results verify that the proposed current-sharing approach is effective.