Abstract: Research into the optimal configuration of distributed power sources can reduce energy costs and environmental pollution; however, their large-scale integration may compromise the stability of the power system. Therefore, an improved Grey Wolf Optimisation Algorithm is proposed, which incorporates the update mechanism from the Particle Swarm Optimisation algorithm to accelerate the update rate of the Grey Wolf’s position, and employs a non-linear strategy to avoid local optima. Simulations on an IEEE-33 node network demonstrate that this algorithm improves convergence speed and accuracy, whilst effectively suppressing excessive voltage amplitudes. As a result, the voltage amplitude at test node 18 remains essentially stable at around 0.95 p.u. over the cycle, significantly improving voltage deviation during peak load periods.