With the continuous advance of the construction of a new power system with renewable energy as the core, the penetration rate of wind power in the grid in various regions of China will increase rapidly. However, large-scale wind farms cannot provide the inertia support, and after the introduction of virtual inertia control, there is also a lack of partitioned and accurate virtual inertia evaluation methods. This paper considers the randomness and correlation of wind speed in a wind farm, and proposes a wind farm partition virtual inertia estimation method based on the Copula function and a clustering algorithm. First, the wake and delay effects of wind speed are considered. The probability distribution model of the wind speed of each turbine in the field is established. Secondly, from the wind speed distribution characteristics of each turbine, a spectral clustering algorithm is used to cluster the turbines. Then, the center turbine is selected in each area. The optimal Copula function is constructed to describe the wind speed correlation between each area. Finally, the partitioning method is used to estimate the virtual inertia reserve of each area in the wind farm. This paper constructs a simulation case based on the actual wind speed and output data of a wind farm in Gansu. The simulation results show that the algorithm proposed in this paper can effectively realize the feature extraction, clustering and division, and inertia estimation of the virtual inertia in the wind farm. This work is supported by the Science and Technology Project of State Grid Corporation of China (No. 522722191005).