To effectively study the dynamics of power systems with large-scale wind farms (WFs), an equivalent model needs to be developed. It is well known that back-toback converters and their controllers are important for the dynamic responses of the wind turbine (WT) under disturbances. However, the detailed structure and parameters of the back-to-back converters and their controllers are usually unknown to power grid operators. Hence, it is difficult to build an accurate equivalent model for the WF using the component model-based equivalent modeling method. In this paper, a transfer function based equivalent modeling method for the WF is proposed. During modeling, the detailed structure and parameters of the WF are not required. The objective of the method is reproducing the output dynamics of the WF under the variation of the wind speed and power grid faults. A decoupled parameter-estimation strategy is also developed to estimate the parameters of the equivalent model. A WF that consists of 16 WTs is used to test the proposed equivalent model. Additionally, the proposed equivalent modeling method is applied to build the equivalent model for a real WF in Northwest China. The effectiveness of the proposed method is validated by the real measurement data.