Artificial Intelligence (AI) is the central driving force behind the new round of technological revolution and industrial transformation. Responsible AI, which focuses on the safe, reliable, and ethical development and use of AI technologies, is of great significance in promoting the healthy development of AI. This paper defines the concept of responsible AI and constructs an analytical framework for the governance system, characterized by “whole-process—multientity—multi-level—multi-mechanism” dimensions. Based on this framework, it analyzes the policy systems of major countries around the world in advancing responsible AI governance, with a particular focus on the AI policy measures in the United States. It examines and summarizes the key initiatives and specific practices adopted by the US in developing responsible AI. The study shows that developed countries represented by the United States, place high importance on responsible AI, involving coordinated efforts from governments, industry organizations, enterprises, research institutions, the general public and individuals. These stakeholders collaborate in governance, utilizing a multi-layered set of tools, including policies, regulations, ethical guidelines, and standard systems. They aim to establish and improve comprehensive governance mechanisms for decision-making and consultation, overall coordination, information flow and sharing, and monitoring and supervision, ensuring the safety, reliability, and controllability of AI technologies throughout their lifecycle. Finally, the paper proposes recommendations for advancing the construction of a responsible AI governance system in China, including strengthening strategic and policy guidance, improving the regulatory system based on the entire innovation process, enhancing the participation of social groups and the public, and fostering international cooperation.