In order to adapt to the grid form after a high-proportion of renewable energy is connected, this paper gives a variety of different power flow margin allocations from the perspective of installed capacity, start-up capacity, ultra-short-term power prediction, power completion rate and minimum abandoned wind and light. Also it compares and analyses their advantages and disadvantages. A power flow margin-weighted distribution method with the smallest abandonment of wind and light is proposed. This comprehensively considers the assessment and scoring indicator of new energy stations and power completion indicators, and proposes a corresponding solution based on an orthogonal learning differential evolution algorithm. The rationality and effectiveness of this power flow margin allocation method are analyzed by simulation of the New England 39-bus system. To help ensure the safety and stability of the power grid, the transmission channel resources of the power grid are used to the greatest extent, and the power grid's ability to accept new energy is maximized. This realizes the dynamic and optimal distribution of real-time output of new energy plants, and ensures the maximum output of new energy plants. This work is supported the National Natural Science Foundation of China (No. 51867005) and the Science and Technology Foundation of Guizhou Power Grid Co., Ltd. (No. 066500KK52170037).