The paper introduces a Fake News Propagation Simulation (FPS) framework based on large language models (LLMs) to study the dynamics of fake news propagation and control. Each agent in the simulation represents an individual with distinct personality traits and is equipped with both short-term and long-term memory, as well as a reflective mechanism to mimic human-like thinking. Agents engage in random opinion exchanges, reflect on their interactions, and update their opinions daily. The simulation reveals patterns in fake news propagation related to topic relevance and individual traits, aligning with real-world observations. The study also evaluates various intervention strategies, demonstrating that early and frequent interventions are most effective in controlling fake news propagation. The research highlights the utility and potential of LLMs in combating fake news, offering valuable insights for practical applications.The paper introduces a Fake News Propagation Simulation (FPS) framework based on large language models (LLMs) to study the dynamics of fake news propagation and control. Each agent in the simulation represents an individual with distinct personality traits and is equipped with both short-term and long-term memory, as well as a reflective mechanism to mimic human-like thinking. Agents engage in random opinion exchanges, reflect on their interactions, and update their opinions daily. The simulation reveals patterns in fake news propagation related to topic relevance and individual traits, aligning with real-world observations. The study also evaluates various intervention strategies, demonstrating that early and frequent interventions are most effective in controlling fake news propagation. The research highlights the utility and potential of LLMs in combating fake news, offering valuable insights for practical applications.