In an intrusion detection system, the existing system uses single sensor that causes false alarms due to lack of accuracy and also leads to misuse of fuels that are stored in nuclear power plants/ reactors. The fuels that are used in the reactors are too expensive and high in price. The fuels use in nuclear power plants/ reactors must be protected under such circumstances that may reduce the detection and inactive secure system of false alarms under given area. In this paper multiple sensors are used to interface with Arduino to overcome false detection using KNN algorithm for the classification of nearest neighboring values that compares with the predefined and predicted values in the given database. Hence the KNN uses to forecast a new point in sample classification, a database where the data points are divided into many groups. This is the point where KNN is located in the Scikit-learning algorithm list. Therefore, the K Nearest Neighbors method stores all available cases and starting lineup cases on the basis of mutual information. © 2021 Elsevier B.V., All rights reserved.