This paper presents a research method of quality analysis and prediction of intelligent watt hour meter in operation. Based on the relevant data of the key links of the electric energy meter, the data of the electric energy meter in designing, material procurement, manufacturing, acceptance testing, running, demolition and scrapping are selected as the sample data for the model construction, and the XGBoost algorithm classification method is used to establish the quality analysis model of the intelligent electric energy meter. Firstly, taking the data of dismantling meter of State Grid Henan electric power company as an example, the modeling, analysis and prediction of various quality problems of intelligent electric energy meter are carried out, and the field verification is carried out, and then the model is continuously optimized according to the verification results. The results show that the accuracy of this method is 0.73, which can objectively reflect the quality of key links of intelligent watt hour meter.