In the context of the new power system, traditional energy consumers are transforming into prosumers who can both consume and produce electricity. The emergence of a large number of prosumers has made the purchase and sale decisions of electricity retailers more complicated, and at the same time, electricity retailers will also face more uncertainties, resulting in greater market risks. In view of the above problems, this paper proposes an optimization method of power purchase and sales strategy for electricity retailers facing prosumers under uncertain factors. Firstly, a two-tier optimization model for the purchase and sale of electricity by electricity retailers considering uncertainties is established. The upper model aims at the comfort and cost of prosumers, and considers the uncertainty of photovoltaic output, and establishes a robust energy optimization model for prosumers. Aiming at the uncertainty of electricity purchase and sale prices in the spot market, the lower model establishes a robust comprehensive decision-making model of electricity sales companies based on info-gap decision theory (IGDT). Then, the Karush-Kuhn-Tucker (KKT) condition is used to transform the two-level optimization problem proposed in this paper into a single-layer nonlinear programming problem. Finally, the simulation analysis proves the economy of the retail electricity price obtained and the effectiveness of established model.