Fe-based soft magnetic composites are widely used in power electronics and power system equipment due to their excellent magnetic properties and low iron loss. As a key component, the performance of the core reactor directly affects the operation efficiency and stability of the power system, and the traditional design method is difficult to take into account the electromagnetic performance and noise control at the same time. In this study, genetic algorithm is used to co-optimize the core structure, electromagnetic parameters and noise characteristics to reduce losses, improve electromagnetic compatibility, and reduce the noise generated during operation. In terms of methodology, a multiphysical field calculation model is constructed based on finite element analysis, electromagnetic performance and noise source characteristics are simulated, and genetic algorithm is used to optimize the parameter combinations under the constraints to form an optimized design scheme. During the optimization process, a suitable objective function is selected and combined with a multi-objective optimization strategy to balance the electromagnetic performance and noise suppression effect. The results show that the optimized core reactor is better than the traditional design in terms of loss, magnetic field distribution and noise level. The optimization scheme derived from the study can effectively improve the electromagnetic characteristics of the equipment and significantly reduce the noise level, providing strong support for the design and improvement of related equipment.