Abstract: Drilling end-effector is the main function integrated equipment for automatic drilling of aircraft components, and its design method has a significant impact on the drilling process. Starting from the part of structural design, based on the neural network fitting algorithm, 80 groups of 5-Dimensional vector sample data including three types of frame components material, displacement and mass are used to train the network model with a mean square value MSE of 0.06, and the key dimensional parameters of the end-effector frame through prediction are obtained, and then the topology optimization method is used to design for lightweight of the frame components. Before topology optimization, the maximum deformation of the frame structure under actual working conditions is 0.04 mm, the maximum stress value is 44 Mpa, and the requirements of the stiffness and strength are met; Under the same working condition, the maximum deformation of the frame structure after topology optimization is 0.026 mm and the maximum stress value is 42 MPa. While the mass is reduced by 43%, the performance of the frame structure of the drilling end-effector is also improved. The prediction of the key dimension parameters of the frame based on the neural network fitting algorithm makes the empirical judgment is no longer simply relied on to design, and the new idea and method for the structural design of the end-effector is developed.