Abstract: The random fluctuation of wind power is the bottleneck of the improvement of wind power prediction accuracy. On the one hand, the volatility of wind speed makes the wind power fluctuate; on the other hand, the ability of a wind farm to convert wind energy into electrical energy will also cause wind power fluctuations to some extent. Therefore,this paper first analyzed the necessity of taking into account the status of wind farms in power forecasting, and then uses the random matrix theory to evaluate the status of wind farms.Based on this, a wind power forecasting method was considered based on the state of wind farms. The results of the example show that this method can effectively improve the ultra-short-term prediction accuracy of wind power.