Abstract: Abstract: This paper takes abnormal audit fees as the proxy of audit collusion and utilizes empirical method to test whether analysts can detect audit collusion. The research has found that the following number of analysts will be significantly reduced with the increase of abnormal audit fees, which indicates that analysts can perceive the risk brought by audit collusion and give the market information by stopping following. Further study has also found that analysts are more sensitive to abnormal audit fees of the companies listed for a long time and the private enterprises, thus resulting in the dramatic decline of analysts following number.