Abstract: The high permeability of distributed generation (DG) units will cause various power quality problems, which will lead to the reduction of power grid carrying capacity. Therefore, a method to improve the load-carrying capacity of DG connected to power grid based on improved grey wolf optimization and gravity search (GWO-GS) is proposed. Firstly, a harmonic filter is proposed to suppress harmonics and reduce power loss, so as to maximize the random carrying capacity of harmonic constraints. Then, the adaptive adjustment of control coefficient and the effect of fitness value on weight are used, and the improved gray wolf algorithm is used to optimize to avoid falling into local optimum. Finally, combined with uncertain factors such as DG intermittent output power and load change, the filter optimization based on improved GWO-GS is constructed. In addition, Matlab simulation platform is used to verify the proposed optimization algorithm. The results show that the proposed algorithm has the shortest convergence time and better performance than other algorithms.