Localization is to determine one’s position in the environment. It benefits a wide range of computer vision applications including autonomous driving, augmented reality, and robot navigation. In recent years, the rapid advancement of end-to-end learning methods has provided novel and promising perspectives for visual localization. Localization based on end-to-end methods has undergone extensive research, aiming to construct robust, reliable, and accurate visual localization architectures. This chapter reviews localization based on end-to-end learning methods, which are divided into two categories: regression-based localization and hybrid localization. This chapter describes and critically evaluates the current research status, advantages, and limitations of existing algorithms in this field. In addition, discussions on datasets, evaluation metrics, and potential future trends are provided, aiming to offering a panoramic theoretical reference and valuable guidance for relevant researchers.