<p>In the field of electric vehicles and new energy, dual-active-bridge (DAB) DC–DC converters face incompatibility between current stress and backflow power under soft switching conditions, which affects the power conversion efficiency. To address this problem, a decomposition-Lagrange multiplier based on the multi-objective evolutionary algorithm (MOEA) is proposed in the paper. First, a fundamental model based on the three-phase-shift modulation is established for the DAB converter. Second, by analyzing the effects of the dead time and soft switching on the power conversion efficiency, the MOEA is used to select weight factors for light, medium, and heavy load conditions. The optimal phase-shift angle for minimal current stress and backflow power is determined. Finally, the results are validated using a small-scale DAB prototype. The experiment showed that the dead time had different effects on the voltage and current waveforms of the DAB converter under various operating conditions. Compared with traditional methods, the proposed method effectively reduced both current stress and backflow power to improve the conversion efficiency.</p>