Man-in-the-Middle (MiTM) attacks pose substantial risks to Internet of Things (IoT), compromising the integrity and confidentiality of data transmitted between nodes. IoT devices are becoming more common, which makes them appealing to attackers. This study investigates the classification of MiTM attacks in IoT networks by utilizing three machine learning (ML) algorithms: (i) Decision Tree (DT), (ii) Random Forest (RF), and (iii) Naîve Bayes (NB). The study employs datasets that include both regular and hostile traffic to train and test the performance of the models. The selected dataset is the IoT Network dataset from the TON IoT (UNSW-IoT20) dataset, with a specific emphasis on MiTM attacks. The results demonstrate that machine learning algorithms may successfully differentiate between legitimate and malicious activity, offering a strong mechanism for improving security in the IoT.