In order to solve the problem of the high-precision kinematics inverse solution modeling of the parallel automatic filling robot, the parallel automatic filling robot is taken as the object, and its kinematic calibration is studied. The pose error model is established by the vector differential method. Based on this, the influence of each error source on the terminal pose of the moving platform is analyzed to find a suitable calibration position. The objective model of the root mean square error is taken as the objective function. The particle swarm optimization algorithm is used to optimize the neural network structure. Finally, it is verified by simulation experiments. The simulation results show that the method can effectively improve the accuracy of the parallel automatic filling robot and provide a theoretical basis for subsequent experimental applications.