Wireless sensor networks (WSN) are increasingly exposed to denial of service (DOS) attacks, which may disrupt network functions and cause significant damage. DOS attacks aim to flood the network with fake traffic, preventing legitimate users from accessing services. The research demonstrates the ability of machine learning techniques to identify wireless sensor network spoofing-of-service (DoS) attacks. Four categories of attacks are detected and classified: Blackhole, Greyhole, Flooding, and Scheduling attacks using a specialized WSN dataset. In this study, two main goals are pursued. Firstly, the focus is on determining the most effective classifier among 17 different classifiers. The second objective aims to apply feature selection method to datasets. The performance of each technology is evaluated based on A set of metrics the results highlight the superiority of the Random Committee, Random Forest and Random Tree classifiers, which showed very strong performance, each achieving an accuracy rate of 99%.