Abstract: To address the problem of significant computational and communication burden, poor robustness and high fault risk when the centralized active power distribution method is applied to large-scale wind farms, a wind farm active power scheduling method based on distributed congestion control is developed. Firstly, considering the mechanical fatigue caused by pitch angle adjustments and the control mode switching due to excessively fast rotor speed reduction, a cost function quantifying the sensitivity of the wind turbine's power increment to pitch angle and rotor speed is introduced. Secondly, a congestion index is designed to optimize power distribution and tracking performance. Finally, the distributed consensus algorithm is used to simplify the calculation of wind turbine power reference, significantly reducing the computational and communication burden on the control center and endowing the proposed method with scalability. A comparative analysis with centralized control methods demonstrates the superiority of the proposed method in power tracking performance, power distribution, and robustness.