Interesting Information is extracted from the data acquire system by data mining technology, a method of smart meter online monitoring based on association rules of data mining is proposed, which implementation mode is designed according to the structure of data acquire system. The structure of online monitoring system is composed by transport layer and application layer. The fault information is produced after analyzing the data of the abnormal operation status of smart meter using real-time monitoring by the transport layer. Association rules are calculated from the database using offline operation mode by the application layer, and the fault diagnosis is realized using online operation mode by the comparison of the rules and fault information after transition. Results of calculation example indicate that the proposed method is effective and accurate for fault diagnosis, and available for application.