It is necessary to accurately predict converter losses for optimized design of a high-power DC-DC converter. The losses of switching devices and inductor among the elements of the converter take significantly greater proportion. The current ripple will be determined by the size of the inductance and this inductance value varies depending on the DC amount of inductor current. As the inductance changes according to load current, the change influences not only the inductor loss itself but also the total converter loss. In this paper, for more accurate design of a bi-directional DC-DC converter for 30kW-class energy storage system, more accurate computational model is proposed considering inductance variation according to the load current change. The inductance changes using variable magnetic cores are verified and converter efficiency is tested through simulations and experiments.