Numerical optimization analysis method is introduced. The workflow suitable for engineering structure design optimization is established based on the optimization software Noesis-Optimus. Corresponding software interfaces and analysis file generation codes are developed to form an automatic design analysis method. The machine learning strategy is used and the correlation analyses are performed on input parameters, constraints and optimization goals. The single-objective and multi-objective optimizations are carried out for structural strength, fatigue life and regional structural weight, respectively. Through the analysis of thousands of samples, multiple sets of optimal design solutions and Pareto fronts are obtained, which can be used as design alternatives. The quantitative automatic iterative optimization is applied to replace the qualitative manual trial calculation optimization as technical breakthrough. While the optimal solutions are obtained, the influence of related design parameters on the optimization goal is grasped and a large number of alternatives are found.