Accurate model of the wind farm is the basis for the analysis of wind power integrating grid. Firstly, the comparison is made between the standard power characteristic curve and the measured scatter plot of wind speed-active power for DFIG. The wind speed-active power model is proposed based on the measured data. For the wind speed differences of wind turbines in large wind farm caused by complex terrain and the crew irregular arrangement, K-means clustering algorithm is used for the equivalent wind speed model of the whole wind farm. Then the wind speed-active power model of wind farm based on measured data is proposed. Taking an actual wind farm as example, K-means clustering algorithm is used to the clusters division. Result shows that the division by wind speed is different from that by location of wind turbine units. Finally, to verify the effectiveness of the new wind speed-active power model, the error comparative analysis is made between the new model and the traditional model. Result shows that the accuracy of the model is greatly improved with the proposed method.