As computer networks play increasingly vital roles in modern society, information security becomes one of the most important research issues in the field of information technology. But intrusions cause a serious security risk, how to efficiently prevent and detect intrusions becomes one of hot research problems in the field of information supervision. The traditional process of building the model of intrusion detection is slow, whose cost of research and development is high. However, data mining has unique advantages in acquiring unknown knowledge. So, intrusion detection based on data mining becomes a hot issue. The research background, architectures, techniques, problems to be solved and the future direction are discussed after analyzing current status of network intrusion and situation of R&D on intrusion detection and data mining.