Abstract: In recent years, with the rapid increase in the proportion of photovoltaic installation capacity, the more and more prominent photovoltaics output volatility and anti-peak regulation characteristics have intensified the pressure on the power system peak regulation. The application of distributed photovoltaics and energy storages resources to participate in the peak shaving has become an effective measure to alleviate the pressure of peak shaving in the power system. However, the high proportion of distributed photovoltaics and energy storages directly participating in peak shaving will lead to the explosion of the dimension of decision variables and the difficult convergence of solution results. In this regard, this paper aggregates the high proportion of distributed photovoltaics and energy storages into a smaller number of feature clusters based on the typical peak shaving characteristics. An aggregation-peak shaving-decomposition model is established, which effectively solves the problem of high proportion distributed photovoltaics and energy storages participating in the power system peak shaving optimization. The results of calculation examples show that the proposed method effectively reduces the dimension of variables and the solution difficulty. The feasibility of its participation in the optimization of power system peak shaving is improved, and it has certain economic advantages while ensuring the peak shaving capability.