Bonding mechanics governs stress transfer in Textile Reinforced Concrete (TRC) through the interface between the cementitious matrix and the textile impregnation material. Epoxy resins are widely used, while mineral impregnation systems (e.g., geopolymers) have been proposed recently as potentially providing stronger and more compatible bonding with cementitious matrices. However, direct, quantitative comparisons of their adhesion to cementitious materials remain limited. Conventional pull-out testing [1] is challenging for this comparison because geopolymers are brittle in tension, whereas epoxy exhibits a ductile response. Push-out tests provide a more consistent basis for assessing interfacial shear resistance for both material classes. Push-out testing is nevertheless challenging because the surrounding cement matrix may fracture under push-out–induced normal stresses. In addition, the true interfacial strength cannot be obtained directly, since the shear stress distribution is non-linear and governed by coupled adhesion and friction [2]. Finite element (FE) analysis is therefore used to optimize the push-out setup to achieve successful interface-dominated push-out and to provide a reliable tool for bond assessment. For this purpose, five push-out configurations are examined, varying the outer matrix (cement or epoxy), the cured core material (epoxy or geopolymer), and the core protrusion (0 or 2 mm). indicating that the governing failure mode is controlled by the relative core–matrix stiffness and by the brittle–ductile material response, through changes in the interfacial stress distribution caused by core protrusion (0 or 2 mm). FE models reproduce the response for each configuration and track the governing failure mode, which depends on the geometric and material parameters. The validated FE model is then used to optimize the push-out setup to promote interface-dominated failure, enabling reliable comparison of cement–epoxy and cement–geopolymer bonding.