Abstract: Clarifying the characteristics and rules of the unsafe behavior of miners is the key content to improve the management effect and ability of the unsafe behavior of miners. This paper proposed a method to analyze the characteristics of miners’ unsafe behavior based on frequency analysis method and association rule algorithm, constructed a process model of miners’ unsafe behavior based on association rules, and obtained the strong association rules and key factors between unsafe behaviors from the result analysis. The results show that the unsafe behaviors of miners are significantly related to the “unit”, “shift” and “work content”, while the unsafe behaviors of major hazards are significantly related to the working position and time period of miners.