There are many drawbacks and inconveniences in the application of human power in the energy meter calibration line. In order to achieve a standardized level of operation and to improve the efficiency and quality of the automated meter testing line, this paper applies the intelligent inspection robot to the automated meter testing line and discusses the key technologies involved. Based on the texture characteristics of the screws on the energy meter cover, a screw coordinate positioning method based on the texture center of gravity method is designed as a machine vision technique for intelligent inspection robots. Based on the feedforward controller transfer function and feedback system open-loop transfer function, combined with PI controller, a feedforward-feedback composite servo position control strategy is designed to complete the release action of the robot end controller. Pressure sensor on the robot end claw controller, integrated servo drive with current sensor. The Kalman filter method of static estimation is used to fuse and process multi-source data information to realize the grasping action of the robot end controller. The test results of the key technical performance and economic and time benefits of the robot show that the recognition success rate and grasping success rate of the robot for energy meters are as high as 100%, and it takes a total of 54s to complete each grasping and releasing action. The maximum error in each direction is 4.9mm, 5.2mm and 5.1mm respectively, and the maximum error in angle is only 1.25 degrees. The working manpower is reduced by as much as 93.16%, the average expenditure of inspection cost is only 1.24 yuan, and the floor space is reduced to 700 square meters. In summary, the study can ensure a high level of consistency in the quality of energy meters, improve the efficiency of calibration and production, and create greater economic benefits while providing a solid technical guarantee for the large-scale construction and stable and reliable operation of the power grid.