With the development of industrial modernization, more and more manipulators are used in industrial production. Standard manipulators are mainly industrial manipulators with high precision and low load, which are mainly used to carry out tasks such as handling, assembly and spraying in the production line. With the rapid development of manipulator technology, outdoor heavy-duty tasks have gradually changed from manual operations to mechanical operations. Large vehicle-mounted manipulators are the main types of manipulators currently used for outdoor heavy-duty tasks. With its large load, large working range, and convenient transportation, the vehicle-mounted large manipulator has a wide range of applications in cargo handling and assembly tasks. However, there are still many problems in the large vehicle-mounted manipulators, such as: low accuracy and low degree of automation. The precision problem of vehicle-mounted manipulator is one of the main problems that plagued its development. How to improve the accuracy of vehicle-mounted large-scale manipulators at a lower cost is the focus of scholars at home and abroad. By improving the accuracy of large-scale manipulators, not only expand the use range of vehicle-mounted large-scale manipulators, but also lay the foundation for the automation development of vehicle-mounted large-scale manipulators. This paper takes the vehicle-mounted manipulator as the research object. The article analyzes the error distribution law of large redundant manipulators and conducts the kinematic calibration study. By analyzing the error change rule of the manipulator, it provides a theoretical basis for the processing and assembly of the manipulator. The manipulator kinematics calibration technology is used to improve the positioning accuracy of the manipulator to meet the accuracy requirements of the manipulator handling task. The main research contents of this paper are as follows. (1) The kinematics modelling of the sizeable redundant manipulator is established. This paper introduces the research object PK1800 manipulator. According to the characteristics of PK1800, the DH model of the manipulator is established to solve the forward kinematics equation of the manipulator. Robotics / Toolbox is used to determine a simulation model of the manipulator and verify its correctness. By introducing the basic principles of mechanical arm differential kinematics, the motion conversion relationship between joint space and Cartesian space is given. (2) The inverse kinematics solution of the sizeable redundant manipulator is obtained. The PK1800 manipulator has moving joints and does not meet the PIEPER criterion (the three adjacent joint axes of the robot intersect at one point or three axes are parallel). According to the characteristics of the PK1800 manipulator, an artificial intelligence algorithm is selected to solve inverse kinematics problem. According to the basic principle of the improved artificial bee colony algorithm, MATLAB is used to address the inverse kinematics of the PK1800 manipulator. The iteration curve of the pose error and the change curve of the error in each direction of the space are obtained, and the effectiveness of the inverse kinematics method is verified by using the spatial motion trajectory. (3) The error analysis of the sizeable redundant manipulator is carried out. The primary error sources of the manipulator are introduced. According to the characteristics of the primary error sources, the error analysis is carried out using the Monte Carlo method. The variation rules of the manipulator posture plane error, zero error and joint sensitivity are obtained. (4) The kinematic calibration method for the sizeable redundant manipulator is carried out. The singularity of the DH model of the manipulator is modified, and the MDH model is established. Based on the characteristics of large redundant manipulators, error models based on position error and distance error were established. Simulation experiments were performed using MATLAB to analyze the relationship between parameter errors and the number of sampling points to obtain the identification results of joint parameters and compare the two the pros and cons of the model. (5) The calibration experiment of the sizeable redundant manipulator is carried out. According to the structural characteristics of the large redundant manipulator, the calibration object is simplified. An error model based on distance error is selected. The principle of binocular distance measurement is used to determine the movement distance of the manipulator in space. The obtained space distance is substituted into the calibration program to identify the geometric parameter error of the manipulator, and then verify the calibration effect of the manipulator. The results show that the accuracy of the manipulator has been improved after kinematic calibration.