Automatic verification assembly line provides guarantee for the normal operation of smart meters. However, in the long-term operation process, performance degradation or even failure may occur, especially the deformation or rust of the mechanical link of the meter positions, resulting in deviation of the error test results. Consequently, the maintenance of meter positions is extremely necessary. At present, the maintenance of meter positions depends on the manual detection carried out regularly, and it is unable to respond to the abnormal working conditions occurring in the maintenance interval in time, and the abnormal meter positions still serve the risk. Therefore, it is of great significance to implement the online detection of the abnormal meter positions in the assembly line and to identify the problems in time. This paper proposes a method for online detection of meter positions’ anomalies in automatic verification assembly line, extracts the characteristics of the verification data distribution of the meter positions, and converts the abnormal state of the meter positions into the abnormality of data distribution. Then, the local outlier factor algorithm is used to quantify the abnormal degree of distribution, and the meter positions with abnormal distribution will be marked. Finally, this paper analyzes the smart meter verification data of the Metrology Center of Shandong Electric Power Company, and the results show that this method is effective.