In the recent decade, the world start moving towards a new era of connectivity where billions of sensors are connected over a network called the internet of thing (IoT). IoT enables a wide range of physical objects and devices to be connected and monitored in fine spatial and temporal detail. Despite their potential of improving multiple application domains, anomalies in the device’s behavior rise a big chal-lenge, especially in smart cities domain. The task of identifying the IoT anomalies in smart cities by visual examination is tedious at best, and more likely, impossi-ble. Moreover, this task is not a one-time effort but rather an ongoing effort. It is, therefore, crucial to find automated ways to help with this challenging task. This paper presents a review of anomaly detection techniques using statistical and ma-chine learning approaches used in smart cities. The paper begins with the explana-tion of essential contexts related to IoT and its architecture followed by a compre-hensive review of IoT anomaly detection and its challenges, types, and detection modes. The paper then presents a review for the related work within smart cities domain. The paper then presents a discussion for the studies being reviewed. Final-ly, the paper highlights the open challenges found in IoT anomaly detection in smart cities domain.