Abstract: For problems that wireless transmission delay in tunnels of coal mine underground was random greatly and current time synchronization algorithms could not be applied directly in coal mine underground because of inability to determine delay distribution and parameters, an estimation method of transmission delay of time synchronization information for Internet of things was proposed based on passive measurement. A class of evaluation node is added to perception layer of Internet of things of coal mine underground for overhearing passively time synchronization packets among perception nodes to acquire transmission delay data. Then parameters under different distribution regularities are estimated by the evaluation nodes based on the maximum likelihood estimation method and optimal distribution regularity is determined according to Kullback-Leibler difference value. Parameter mutation is detected by using log-likelihood-ratio for determining whether the evaluation nodes send delay distribution regularity and parameters to sink nodes. The simulation results show that the method detects mutation of classification of delay distribution regularity and distribution parameters accurately.