The problem of intrusion detection into a corporate network is considered in this article. The main components of intrusion detection system are allocated and their functions are described. Various approaches to identification of violations of information security are analysed. For this purpose, the characteristic of the main methods of intrusion detection is given, their merits and demerits are allocated. It is shown that to increase the efficiency of detection of possible invasion situations, it is necessary to use advanced data mining technologies. So for the purpose of application in intrusion detection the features of the data mining technologies are investigated, by the results of their comparative analysis hybrid means for identification of attacks are offered. It is shown that is the most perspective for a considered task use of neuro-fuzzy methods. The architecture of neuro-fuzzy system for intrusion detection into a network is offered.