Federated Edge Learning (FEEL) emerges as a promising approach in distributed machine learning by leveraging the computational power of edge devices to train a global model in a distributed fashion. This chapter provides an overview of FEEL, highlighting its fundamental concepts, architecture, and operational procedures. The chapter begins by introducing the principles of FEEL, and then delves into the various basic machine learning models and algorithms employed in FEEL, ranging from first-order, second-order, and zeroth-order algorithms. Furthermore, we discuss FEEL from the perspectives of algorithm efficiency, network architecture, and privacy and security, while also outlining the challenges inherent in implementing FEEL.