Abstract: The most important element in a network system is to preserve the security of vulnerable networks; numerousarchitectures have been proposed for network protection or to eliminate unwanted access to internal and externalconnections. Various existing systems have developed diverse approaches to detect suspicious attacks on victims devices;nonetheless, an external user builds harmful behaviour and gains unauthorised access to victim workstations through such abehaviour framework, known as malicious activity or Intruder. To discern between them, a range of supervised machinealgorithms and soft computing methods have been createdReal-time events as well as simulated network log data The NLSKDD data set is the most widely utilised data set to identifythe Intruder on the benchmark data set. In this research, we propose that intruders be identified using machine learningmethods. Two related strategies have been proposed: signature detection and anomaly detection. The Recurrent NeuralNetwork (RNN) algorithm is presented with various data sets in the experimental study, and the system's output isdemonstrated in a real-time network setting.Keywords: , Recurrent Neural Network, KDDCUP99, Intrusion Detection System, Network security