Mobile IoT unleashes an extraordinary cycle of innovative applications that target humans, vehicles, and mobile objects, e.g., intelligent asset management. These applications necessitate the smart recognition of objects and their attributes, the real-time tracking of the location and status of things, the on-time understanding of heterogeneous contexts, etc. This tutorial focuses on sensing, recognition, and tracking of humans and objects for mobile IoT applications, which are key enablers for cyber-physical systems and digital twin concepts. The tutorial consists of four parts. First, as for cost-effective and maintenance-free context recognition, we review wireless sensing leveraging RF signal change such as Wi-Fi to recognize various contexts in intelligent spaces. Second, point clouds from LiDAR and depth cameras are also widely used for a precise understanding of 3D physical space. We will review basic point cloud processing, deep learning-based spatial recognition, and applications such as large-scale human tracking in crowded environments. Third, since location is one of the most valuable user-context information, we will discuss the current trends in localization techniques powered by the recent advancement in artificial intelligence (AI). This tutorial will also include the current trends in energy-free location determination technologies. Finally, we show smart healthcare technologies based on sensing data and machine learning for health monitoring to solve the health problem, which is one of the most significant issues worldwide.