Metamodeling employs simplified surrogate models to predict the behavior of complex structures at a fraction of the computational cost of full-scale simulations [1]. This approach enables efficient exploration of large design spaces and identification of optimal configurations [2]. In the present study, a sandwich structure with an auxetic core subjected to low-velocity impact is investigated and optimized using a metamodeling framework. The numerical model was validated against experimental data available in the literature, ensuring reliability of the simulations. An artificial neural network surrogate model was then trained to capture the nonlinear impact responses and guide the optimization process. The optimized structure demonstrated a remarkable 112.57% improvement in specific absorbed energy compared to the baseline design. Phenomenological analysis revealed that plastic deformation dominated the impact response, with the enhanced energy absorption attributed to a characteristic layered buckling pattern inherent to the auxetic core design. Notably, this favorable deformation mode was not known prior to the study, underscoring the ability of metamodeling to uncover unexpected yet beneficial structural behaviors. By reducing computational expense while enabling discovery of optimal designs, this work demonstrates that metamodeling-driven structural optimization can significantly enhance the impact performance of auxetic materials, offering a pathway toward more efficient and resilient engineering solutions. REFERENCES [1] Q. He, L. Li, X. Jing, Y. Jiang, D. Yan, Impact resistance analysis and multi-objective optimization of polyurea-coated auxetic honeycomb sandwich panels, Materials Today Communications vol 35, 2023, 105577. [2] Costa, E. A.; Driemeier, L. Parametric optimization framework for designing sandwich panels with auxetic core subjected to impact load. Composite Structures, v. 347, art. 118436, Elsevier, 2024.