Abstract: In the context of the dual-carbon goal, virtual power plants, microgrids and other forms of prosumers have emerged on a large scale. Energy sharing among multiple prosumers can improve the overall economic efficiency and new energy consumption. A collaborative operation mechanism for energy sharing among multiple prosumers under the weak centralization model is proposed. The mechanism considers the data privacy of each prosumer. Firstly, an internal scheduling model of prosumers with multiple distributed resources is constructed. The model considers the volatility and randomness of load demand and new energy output, quantifying the risk of uncertainty based on conditional value at risk (CVaR). Then, a weakly centralized tariff iteration mechanism for multiple prosumers is proposed, which uses the supply and demand relationship to guide the tariff update. Meanwhile, considering the privacy protection of prosumers, the power data aggregation method is designed based on Paillier homomorphic encryption algorithm and secret sharing principle. The method is able to obtain the system supply and demand information while ensuring the privacy of all parties. Finally, the effectiveness and rationality of the mechanism proposed in this paper is verified by an arithmetic example, and the overall cost of multiple prosumers is reduced by 12.6% after energy sharing.