<p>This paper represents the multi-objective optimization of a crane hook in such a way that the enhancement of structural performance will be achieved with guaranteed safety and durability. The research study applies a two-phase approach: the first phase, by using validation, a hybrid of Genetic Algorithms (GA), fmincon, and Gray Wolf Optimizer (GWO) (analytically) for the proper minimization of mass in a crane hook while enhancing its structural integrity. In the second case study, a hybrid optimization method of GA combined with fmincon will be applied, followed by topology (numerical) optimization. The proposed hybrid approach with GA-fmincon, along with topology optimization, has allowed a substantial reduction in mass by 26.91% without compromising the mechanical robustness of the crane hook. Such a reduction represents an important achievement over previous works. The improved design exhibits better distribution of stresses, an appropriate factor of safety, and increased fatigue resistance. These improvements are also corroborated by detailed analysis that shows the designed component can resist operational conditions with extreme changes and can also support a large number of loading cycles. A novel convolutional neural network (CNN) is designed, tested, and trained to predict the fatigue life (<i>N</i>) of the crane hook under repeated loading in order to save the time and avoid the complexity of repetition implementation finite element model (FEM). The CNN training phase produces promising results, showing the model performance with 94.79% accuracy, 92.34% regression, and 91.16% F-score at epoch 1000, which shows the effectiveness of the proposed approach. This work approves the optimization methodologies adopted and offers a sound approach that can be further applied in structural engineering applications, with the aim of making the most of advanced computational techniques to achieve material efficiency and structural performance.</p>