Wind power output is closely related to operational conditions. As an extreme weather event

the cold wave often leads to a sharp increase in electricity load. In this scenario

the wind power prediction results may deviate significantly from actual output

posing challenges to the safe operation of the power system. This paper proposes a wind power prediction method for extreme cold wave weather conditions

utilizing RTSGAN sample expansion and attention dynamic weight integration. Firstly

to address the issue of scarce wind power samples during cold wave conditions

RTSGAN is used to expand the original sample data for the cold wave period. Then

the basic prediction models and attention network are constructed based on the expanded sample dataset. Finally

the wind power prediction is carried out by the attention dynamic weight integration method. Experimental results show that the proposed method can address the scarcity of wind power samples during cold wave extreme weather and effectively improve the accuracy of wind power prediction under these conditions.