This paper studies accurate intrusion detection of the multiple mutation networks. The network intrusion viruses may mutate multiple times in a short period, it makes the network intrusion features deviate from the original features much and mutate dramatically. The traditional algorithm detects intrusion based on the original features of the network intrusion. The dramatic changes of the original features will reduce the accuracy rate of the network intrusion. This paper proposes a detection method for multiple mutation network intrusion based on the artificial immune algorithm. This method e...