In wireless charging systems, variations in load parameters can significantly impact the transmission power and efficiency of the system. To enhance the stability and reliability of wireless charging systems, this paper proposes the optimization of Proportional-Integral (PI) controller parameters using an improved Particle Swarm Optimization (PSO) algorithm. Addressing the issue of traditional PSO algorithms easily falling into local optima, this study introduces a strategy of variable inertia weights. Furthermore, to overcome the limitation of traditional PSO algorithms that rely on a single performance metric, which may not accurately reflect the system’s performance, a fitness function setting method considering three performance indicators is proposed to improve system performance. The feasibility of the proposed method is ultimately verified through simulation. The final simulation verification results indicate that the proposed method significantly outperforms traditional PI control. Specifically, during the startup phase, the overshoot is reduced from 12 to 3%, and the dynamic process is smooth and oscillation-free. When a load step change occurs, the recovery time is shortened to 0.003 s, while the voltage and current harmonic distortion rates remain below 1%. These results demonstrate excellent dynamic performance and stability.