Abstract: The reactive power of the doubly-fed induction generator（DFIG） is adjustable within a certain range, so the DFIG wind farm can be used as a new reactive power resource to provide necessary reactive power support to power grids. Due to the nonlinearity of network constraints in wind farms and the uncertainty of the active power of DFIGs, it is difficult to accurately estimate the reactive power support range of DFIG wind farms. A robust estimation method is proposed to solve this problem.Firstly, an estimation model of reactive power support range based on two-stage robust optimization is established. The first-stage optimization gives the candidate maximum reactive power support range, and the second-stage optimization checks whether the wind farm could provide the reactive power in the candidate range in different uncertain scenarios. To facilitate the solution, the nonconvex second-stage problems are transformed into convex optimization problems by cone relaxation and dual transformation,and the accuracy of the relaxation solution is guaranteed by the penalty factor. Then the robust optimization model is solved based on the column-and-constraint generation algorithm to obtain the maximum reactive power support range. Finally, the strategy of the wind farm as a reactive prosumer is given to participate in the reactive power optimization of the power grid. The case study shows that the proposed method can accurately estimate the reactive power support range of the wind farm, and the proposed strategy is feasible.