Abstract: With increasing Internet connectivity and traffic volume, recent intrusion incidents have reemphasized the importance of network Intrusion Detection System (IDS). According to the deficiency of the Naive Bayesian (NB) algorithm, this paper presents an improved NB algorithm, which is addition of an attribute-added method to the traditional NB. This algorithm which is based on the original model and combined with a controlling parameter to enhance the accuracy of classification, the best parameter obtained by experiments can not only simplify both the time complexity and space complexity of the intrusion detection but also optimize the classification performance. The experimental results prove that this proposed approach applied into the intrusion detection framework can drastically reduce the false alarm rate of IDS so as to improve the detection efficiency and decrease economic damage brought by the cyber attack.