With the emerging deployment of 5G wireless systems and the rapid growing number of smart sensors and deploying sensors on or around physical objects, the Internet of Things (IoT) seamlessly integrates a world of networked smart objects, makes their information be shared on a global scale, and provides an ability of intelligent computing and information processing, such as reporting status, position, and surrounding condition of each sensor node. Localization and tracking is a key problem, which has been already studied in various fields, including sonar, radar, seismic, mobile communications, wireless sensor networks. However, many solutions may not directly suit an IoT scenario where large quantities of sensor nodes that perform distributed sensing and collaborative information processing tasks are interconnected together over a wireless channel. It is almost impossible to collect full network sampling data for accurate localization since any inter-sensor communication requires a large burden on sensor batteries. Typical metrics are measured at the local sensors including sample covariance matrices (SCM), time of arrival (TOA), gain ratios of arrival (GROA), angles of arrival (AOA) and frequency differences of arrival (FDOA). Estimating the source position as accurate as possible by utilizing the above-mentioned metrics is full of challenges. In this talk, we introduce the fundamentals of typical source localization and target tracking methods, including least-squares, maximum likelihood, convex relax optimization. We discuss state-of-the-art localization and tracking approaches using global navigation satellite systems (GNSSs), array sensor networks (ASN) as well as for indoor positioning for IoT applications, for example, subband information fusion, auxiliary variables-based algorithms, localization penalized maximum likelihood, and weighted least-squares using AOA-GROA-TOA, and other state-of-the-arts in localization and tracking systems for IoT applications.