This study presents a comprehensive thermal buckling analysis of sandwich plates composed of functionally graded graphene origami-enabled auxetic metamaterial (FG-GOEAM) face sheets on an auxetic concrete foundation, using Carrera’s unified formulation (CUF) as the theoretical framework. FG-GOEAM materials are emerging as advanced composites, combining exceptional mechanical resilience, tunable auxetic behavior, and high thermal stability, making them suitable for extreme environments. By employing CUF, a powerful and adaptable modeling approach, this work accurately captures the complex mechanical interactions within the FG-GOEAM sandwich structure under thermal loads, incorporating both material gradation and auxetic properties. To further enhance the precision and efficiency of this thermal buckling analysis, a deep neural network (DNN) is developed as a machine learning algorithm to predict critical temperature differences, based on a dataset generated through mathematics simulation. The DNN model demonstrates excellent predictive capability, validated by close alignment between its estimates and CUF results, thus reducing computational costs while maintaining high accuracy. Parametric studies are conducted to assess the effects of material gradation, aspect ratios, and foundation properties on thermal buckling performance. The results highlight the superior thermal stability of FG-GOEAM structures and the potential of DNNs to serve as reliable, computationally efficient tools for advanced structural analysis. This study provides a novel, integrated framework for high-fidelity thermal buckling prediction in complex auxetic composites, paving the way for broader applications in engineering fields requiring lightweight, thermally stable structures.