Based on the concept of automatic operator（AO） proposed by Dr. Dy-Liacco， the principle and application mode of AO fitted for China are clarified. Technologies for data mining of fine rules for power system operation is presented， i.e. fine rules for power system security can be discovered automatically through several steps of data mining process such as topic definition， extra large amount of training samples generation， feature selection， rule generation， and the like. A prototype system of AO is then developed， and its PC-cluster based hardware configuration and software structure are introduced. Functions and algorithms adopted in the prototype system are presented in details. Finally， performance tests are carried out for a provincial power system and a case of data mining of a fine rule is given， thus verifying the feasibility and efficiency of online data mining. The fine rule discovered is more suitable for power system operation.