This 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. The FPS framework models individuals as LLM agents with distinct personalities, short-term and long-term memory, and reflective reasoning. Each agent engages in daily random opinion exchanges, updates their beliefs based on interactions, and reflects on their thinking. The simulation captures the evolution of individual and collective opinions, revealing patterns in fake news propagation related to topic relevance and individual traits. The framework also incorporates an official agent to simulate interventions against fake news, demonstrating that early and frequent interventions are more effective in controlling its spread. The FPS framework is validated through comprehensive simulations, showing that political fake news spreads faster than other topics like terrorism, science, or financial information, consistent with prior studies. The simulation also reveals that individuals with certain personality traits, such as high agreeableness and high neuroticism, are more susceptible to believing fake news. The framework's ability to simulate detailed opinion dynamics and incorporate realistic human reasoning makes it a valuable tool for studying fake news propagation and intervention strategies. The FPS framework offers several advantages, including the ability to simulate diverse behavioral patterns, replicate the textual nature of fake news, and provide insights into the dynamics of opinion changes. It also enables the analysis of fake news propagation across different scenarios and demographic groups, offering extensive and valuable insights. The framework's integration of long-term and short-term memory, along with reasoning processes, allows for a more nuanced understanding of how individuals form and change their opinions in response to fake news. The simulation results align with real-world observations and provide practical strategies for monitoring and controlling fake news in real-world scenarios.This 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. The FPS framework models individuals as LLM agents with distinct personalities, short-term and long-term memory, and reflective reasoning. Each agent engages in daily random opinion exchanges, updates their beliefs based on interactions, and reflects on their thinking. The simulation captures the evolution of individual and collective opinions, revealing patterns in fake news propagation related to topic relevance and individual traits. The framework also incorporates an official agent to simulate interventions against fake news, demonstrating that early and frequent interventions are more effective in controlling its spread. The FPS framework is validated through comprehensive simulations, showing that political fake news spreads faster than other topics like terrorism, science, or financial information, consistent with prior studies. The simulation also reveals that individuals with certain personality traits, such as high agreeableness and high neuroticism, are more susceptible to believing fake news. The framework's ability to simulate detailed opinion dynamics and incorporate realistic human reasoning makes it a valuable tool for studying fake news propagation and intervention strategies. The FPS framework offers several advantages, including the ability to simulate diverse behavioral patterns, replicate the textual nature of fake news, and provide insights into the dynamics of opinion changes. It also enables the analysis of fake news propagation across different scenarios and demographic groups, offering extensive and valuable insights. The framework's integration of long-term and short-term memory, along with reasoning processes, allows for a more nuanced understanding of how individuals form and change their opinions in response to fake news. The simulation results align with real-world observations and provide practical strategies for monitoring and controlling fake news in real-world scenarios.