The burgeoning field of Large Language Models (LLMs) presents a novel paradigm in autonomous Agents’ development, heralding significant impacts across diverse sectors. This interdisciplinary study synthesizes the latest advancements in LLM-based Agents, spotlighting their pivotal role in fostering sustainable development, enhancing capacity building, and revolutionizing the creative industry. Combining perspectives from computer science, we have thoroughly reviewed the technical development and cutting-edge progress of Agent-Based Modeling and LLM-based Agents, particularly highlighting their applications and potential in the fields of social simulation and computational social science, underscoring their potential in addressing critical global challenges, including environmental sustainability, health, and education, thereby accelerating the achievement of Sustainable Development Goals (SDGs). Our interdisciplinary team consists of individuals from China and Japan with expertise in computer science, international law, international development, film production, new media, and the internet industry. Our exploration extends to the multifaceted applications of LLM-based agents in sustainable governance and management, particularly explored dynamic individual and group behaviors in both synthetic and physical environments, highlighting agent-based modeling’s capacity to simulate complex social-environmental interactions and inform policy-making for sustainable outcomes. We further investigate the integration of multimodal perceptions, leveraging agents’ abilities to serve as reasoning and planning cognitive core, and process and generate text, image, video, and audio in the context of addressing global challenges in environmental health, climate change, and labor. We analyze the cognitive and behavioral learning systems demonstrating LLM agents’ utility in sustainable negotiation and decision-making processes. The discussion extends to the regulatory frameworks essential for managing external tools, data access, and actions taken by autonomous agents, aligning with digital law and global governance to ensure ethical and equitable deployment of AI Agents. Furthermore, we explore the risks and precautions against the misuse of intelligent autonomous weapons, emphasizing the pivotal role of international NGOs in AI security governance. Furthermore, we delve into the transformative use of LLM-based agents in the creative industries, particularly in multimodal storytelling, where advanced text-to-image and text-to-video generation techniques are applied, showcasing the profound impacts on film and television production in the context of designing sustainable solutions for media and information literacy. In conclusion, this study articulates the transformative potential of LLM-based agents across sectors, advocating for a multi-stakeholder dialogue to harness AI’s power responsibly and inclusively. By bridging technological innovation with ethical governance and creative applications, LLM-based agents stand as a cornerstone in the pursuit of a sustainable future, embodying the essence of “AI for All.”