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.