Abstract: Accidents in coal mines have characteristics of danger sources coupling. The paper applied association rules mining method to study danger source coupling quantitatively. Firstly, the paper looked hidden danger identification as shopping in the supermarket, and carried out modeling analysis, and then the paper carried out "transaction" definition of hidden perils and the analysis of hazards "set", built a Boolean relational table, and established association rules mining model with lift-enhanced Apriori algorithm. Finally, an empirical analysis was carried out for historical hidden danger data in Tiefeng Coal Mine. The results show that the implicit coupling relationship can be effectively extracted by association rules mining method, which provides guidance for safety management of coal mines.