The article reviews the development status and basic principles of federated learning (FL) in the context of edge-Internet of Things (edge-IoT). It highlights how FL addresses the challenges of training complex models with large, heterogeneous, and private data in distributed edge networks. The combination of edge computing (EC) and FL optimizes computing, communication, and data security. The paper discusses various cryptographic technologies (secure multi-party computation, homomorphic encryption, and secret sharing), perturbation schemes (differential privacy), and adversarial training methods to enhance privacy and security. It also addresses the security attacks and privacy leakage issues in edge-IoT and outlines future research directions for integrating EC and FL. The authors emphasize the importance of FL in reducing communication overheads, protecting user privacy, and enabling model training on edge devices, making it a promising solution for the edge-IoT environment.The article reviews the development status and basic principles of federated learning (FL) in the context of edge-Internet of Things (edge-IoT). It highlights how FL addresses the challenges of training complex models with large, heterogeneous, and private data in distributed edge networks. The combination of edge computing (EC) and FL optimizes computing, communication, and data security. The paper discusses various cryptographic technologies (secure multi-party computation, homomorphic encryption, and secret sharing), perturbation schemes (differential privacy), and adversarial training methods to enhance privacy and security. It also addresses the security attacks and privacy leakage issues in edge-IoT and outlines future research directions for integrating EC and FL. The authors emphasize the importance of FL in reducing communication overheads, protecting user privacy, and enabling model training on edge devices, making it a promising solution for the edge-IoT environment.