Abstract: The demand response (DR) of decentralized resources is an essential form of supporting the flexible regulation of the power system. However, the resources typically have a large scale and small individual capacity, necessitating the urgent involvement of market entities such as virtual power plant (VPP) to aggregate, thereby indirectly participating in the regulation services of grid companies. This paper concentrates on the emerging response model of customer directrix load, focusing on the profit-risk interaction and coordination strategy of decentralized resources within VPP, proposing a two-stage optimization and profit sharing-risk sharing decision-making method with customer directrix load, and constructs a new interactive coordination mode between VPP, grid, and internal users. Firstly, under customer directrix load incentive, considering the responsiveness of flexible loads and the accuracy factor of renewable energy, VPP conducts day-ahead and intraday DR optimization to maximize profits. Secondly, under the profit-sharing-risk-sharing decision-making method, VPP shares the marginal incentive with flexible loads and the uncertainty risk with renewable energy, then uses the improved Shapley method to allocate profit to internal loads. The simulation results show that compared to traditional DR, the customer directrix load-based DR provides a superior advantage by offering response objectives throughout the entire period. The sharing decision-making attracts users to participate in VPP optimization and promotes more accurate reporting of new energy in the future.