Abstract: The development of digital technologies has generated regulatory dilemmas, highlighting the growing importance of soft law in the governance of emerging technologies. In response to the development and application of artificial intelligence (AI), a number of countries and regions have adopted regulatory measures, many of which take the form of soft law instruments. At the same time, the formulation of international soft law has gradually become an important pathway for global AI governance. Compared with hard law, soft law governance demonstrates several notable advantages, including the timely clarification of AI-related risks, flexible responses to technological challenges, the provision of a relatively accommodating environment for innovation and development, reduced costs of rule-making, enhanced participation by diverse stakeholders, and the promotion of international exchange and cooperation. However, soft law generally lacks binding enforceability, which makes it difficult to ensure effective implementation. Due to differences in governance philosophies, economies have developed distinct AI governance models based on their respective national conditions, each exhibiting its own strengths and limitations. The number of international soft law instruments on AI continues to increase, yet the overall landscape remains fragmented and characterized by competition. On the basis of drawing lessons from foreign experiences, and guided by the principle of giving equal weight to development and security, China should pursue an AI governance approach that combines soft law with hard law and establishes a dynamic and coordinated relationship between the two. At the same time, it is essential to actively participate in international AI governance, continuously promote the formation of international consensus, and guide AI technologies toward responsible development.