Internet has become a vital part of any organization. But with the growth of internet, intrusion and attacks have also increased. Thus, there arises a need of robust and powerful intrusion detection systems which can detect the attacks. Recently, many novel methods are experimented to build strong IDSs. In this paper, we implement a new Improved Dynamic FCM algorithm and successfully integrate it in WEKA to expand the system functions of the open-source platform, so that users can directly call the FCM algorithm to do fuzzy clustering analysis. Besides, considering the shortcoming of the classical FCM algorithm in selecting the initial cluster centers, we represent this Improved DFCM algorithm which adopts a new strategy to optimize the selection of original cluster centers. A novel classification via dynamic fuzzy c means clustering algorithm has been proposed to build an efficient anomaly based network intrusion detection model. A subset of KDD Cup 99 intrusion detection benchmark dataset has been used for the experiment. The proposed novel concept will be efficient in terms of detection accuracy, low false positive rate in comparison to the other existing methods.