Since the mobile devices has increasingly popular, Millions of specialized applications are already available for variety of platform all over the world. The complexity of this applications often relies on security. The wide variety of networking technologies and devices present challenge when developing efficient strategies to test security of different networking applications. Intrusion in a mobile ad hoc network (MANET) is a hard take problem in today’s network. MANETs are highly susceptible to various dictionary traffics attack due to their inherent characteristics. Intruders can explore vulnerabilities of a network by retrieve unauthorized data through aggregation and inference in order to deploy some malware like dictionary attacks. In this work, frequency-based intrusion detection system (IDS) is proposed for MANETs to detect dictionary node attacks. Frequency data is mined from time serious data created by the flow of traffic by using Discrete Fourier Transform is used to change time series data into frequency data. The frequency of abnormal traffic data would be varying with the frequency of the normal traffic data due dictionary attacks in the network. Frequency-based intrusion detection is used to detect known and unknown attack signatures in a network for detecting dictionary attacks for ad hoc network. The main reason is detecting odd traffic like dictionary attacks at some stage in data passing from one node to the other. Additionally, it is to differentiate the unknown and regarded traffic. This method is tested to shown by running artificial network intrusion data in simulated networks by using the NS2 to get maximum throughput by isolate the abnormal node.