An embodiment of the invention discloses a method for mining an action frequency of a device in reactive power optimization online control, and relates to the field of the reactive power optimization online control of a power system. For the problems of relatively low mining efficiency and difficulty in realizing online setting during calculation by adopting an association mining method in reactive power optimization control, quick and reasonable setting of the action frequency of the device in the reactive power optimization is realized by combining improved quick fuzzy association mining with incremental mining; the overall calculation efficiency of mining the action frequency of the device is improved through the improved quick fuzzy association mining algorithm; and the introduction of the incremental mining further improves the target mining efficiency. Therefore, numerous tedious work of operators in the aspect of setting the action frequency of the device in the reactive power optimization online control is reduced; quick mining and online setting of the action frequency of the device are realized on the premise of ensuring a reasonable and accurate setting result; and a relatively good voltage control effect can be achieved finally in the reactive power optimization online control.