在政治经济动荡、气候变化加剧、技术革新加速的全球化背景下，信息传播生态日益复杂化，虚假信息的泛滥已成为威胁社会稳定、公共健康与政治安全的紧迫问题。文章聚焦大模型驱动的智能体在事实核查领域的创新应用，以期解决复杂场景下的虚假信息治理难题。首先系统梳理了当前事实核查领域面临的挑战，论证了大模型驱动的多智能体赋能事实核查工作的优势。其次设计了基于大模型的事实核查多智能体系统理论框架，通过实证研究指出大模型驱动的事实核查多智能体协同工作的实践路径，并通过对比实验验证了该系统性能优越性。

Against the backdrop of globalization marked by political and economic turbulence, intensifying climate change, and accelerated technological innovation, the information dissemination ecosystem has grown increasingly complex. The proliferation of disinformation has emerged as an urgent issue threatening social stability, public health, and political security. First of all, This paper focuses on the innovative application of large language model-driven agents in the field of fact-checking to address the challenges of disinformation governance in complex scenarios. In the next place, it systematically examines the current challenges faced in fact-checking and demonstrates the potential and advantages of empowering fact-checking work through large language model-driven multi-agent. This research designed the theoretical framework of the large language model-driven fact-checking multi-agent system, outlined a practical path for collaborative work among large language model-driven fact-checking multi-agent through empirical research, and verified the superior performance of the system through comparative experiments.