Abstract: The feasibility of applying artificial immune theory in intrusion detection system is Analyzed, establishes a model combining artificial immune theory and data mining technique is estoblished. Based on the statistical theory, the amount of information that is lost by splitting a data stream into unordered strings can be estimated, and this estimate can be used to guide the choice of string length. Based on information-theory, a lower bound on the size of the detector set is derived. Detector Generating algorithm is described. The performance of Artificial Immune Intrusion Detection System (AⅡDS) is better than the normal intrusion detection system based on knowledge engineering.