Abstract: The structural performance of the column is the key to the stability of the machine tool. In order to improve the dynamic and static characteristics of the column while reducing its mass, we start from the acquisition of the machine structure model, collect the point cloud data of the machine parts through 3D scanning, and then convert it into a solid model by using the inverse modelling technology, and then use the finite element simulation to solve the key indexes of the column and complete the optimization. In terms of model acquisition, 2D analysis is used to compare the deviation between the solid profile of the column and the point cloud data, and the overall average deviation is measured to be −0.050 9 mm, which meets the test requirements. In terms of structural optimisation, multi-objective topology optimisation is used to quickly determine the overall layout of the column structure, and further dimensional optimisation is used to reduce the impact of increased total deformation on the machining accuracy of the machine. Under the condition that the total deformation is basically unchanged, the 1st-order intrinsic frequency of the column is effectively increased by 16%, and the mass is reduced by 5.3%. This method integrates the use of reverse engineering and finite element simulation, provides model parameters obtained in reverse for finite element analysis, increases the engineering reliability of structural optimisation, provides new ideas for product improvement and redesign, accelerates the efficiency of new product development, and saves design and manufacturing costs.