The omnidirectional collection of IP flow information is helpful for the real-time monitoring and precise management of the network,and also beneficial for improving the network security.The original IP flow sampling algorithm makes repeated sampling to the IP flow,resulting in a waste of storage and computing resource.Based on the original algorithm,a new method of IP flow sampling in high speed network with equal probability random is presented.This algorithm adopts three-level bit fields based on Bloom Filter,with two of them measured simultaneously for obtaining the intersection of their results.It is convenient and can prevent repeat sampling to IP flow effectively.The experimental results indicate that:The new algorithm can improve the measure precision drastically and make full use of the system resources,which is applicable to 10 Gb/s high speed network.