Abstract: The randomness of wind power output makes it difficult to balance the robustness and computational efficiency of microgrid grid connection planning. A wind power output scenario construction method matching with typical daily load scenarios is proposed. The daily load trend and the location of peak and valley periods should be considered in the microgrid planning. The daily load curve trend and peak and valley period information are extracted by using the membership function, and combined with the improved ordered clustering, a typical daily load selection method is proposed; In the effective time of typical daily load, using the maximum increase and decrease of wind power output, combined with interpolation method, a wind power scene construction method is proposed. Then an evaluation index system is established to evaluate the selection of typical daily load and the construction effect of corresponding wind power scenarios. Finally, the effectiveness of the proposed model is verified by power grid data.