Comparing with conventional power plants, wind power generations are highly dependent on the natural environment. Especially under extreme wind condition, individual wind turbine might automatically disconnect from the grid once the wind speed is near or over the cut-out speed. This results in the output power reduction of a wind farm, which raises the security concern due to the imbalance between generation and demand, particularly for power systems with high wind power penetration. Different from the continuous wind power fluctuation under normal weather condition, the power output of individual wind turbine drops suddenly under extreme weather. Thus, the power reduction of individual wind turbine requires to be simulated by a higher resolution tool in order to illustrate the sudden power drop under an extreme condition. Also, the wind turbines are dispersed geographically within a wind farm and a region, which has a large impact on the magnitude of power reduction of a wind farm and the region under extreme weather. Therefore under an extreme weather condition, the simulation of wind power reduction requires a model with higher resolution as well as considering "geography dispersion", which cannot be satisfied properly by most of current existing models. This paper hence proposes to use a model in frequency domain to assess the power reduction of wind farms under extremely high wind speed condition. Though this model is originally designed for application under normal wind condition, the originality of this paper is the study of the applicability of this model under an extreme wind condition, because this model provides a second-by-second simulation resolution and uses the coherence matrix in frequency domain to describe the coherences of power reduction among wind turbines. Also, for a regional wind farm cluster, an additional advantage of this model is to provide a reasonable estimate of wind power reduction under extreme wind condition without using extensive history data of the whole region. This model is verified by both qualitative and quantitative analysis and then some statistics-based tools are further developed to assess the reserve requirement due to wind power reduction under extreme wind condition. A case study of Zhangjiakou (ZJK) power region shows the effectiveness of the proposed assessment methodology in an extended wind power system region, and then this model is demonstrated to be valuable for both power system planning and operation with high wind penetration under extreme wind condition. (C) 2011 Elsevier Ltd. All rights reserved.