Purpose: This study evaluated the repair bond strength of differently surface-conditioned press-on-metal ceramic to repair composites and determined the location of the accumulated stresses by finite element analysis. Materials and Methods: Press-on-metal ceramic disks (IPS InLine PoM, Ivoclar Vivadent) (N = 45, diameter: 3 mm, height: 2 mm) were randomly divided into 3 groups (n = 15 per group) and conditioned with one of the following methods: 9.5% hydrofluoric acid (HF) (Porcelain etch), tribochemical silica coating (TS) (CoJet), and an unconditioned group acted as the control (C). Each group was divided into three subgroups depending on the repair composite resins: a) Arabesk Top (V, a microhybrid; VOCO), b) Filtek Z250 (F, a hybrid;3M ESPE); c) Tetric EvoCeram (T, a nanohybrid; Ivoclar Vivadent) (n = 5 per subgroup). Repair composites disks (diameter: 1 mm, height: 1 mm) were photopolymerized on each ceramic block. Microshear bond strength (MSB) tests were performed (1 mm/min) and the obtained data were statistically analyzed using 2-way ANOVA and Tukey's post-hoc test (α = 0.05). Failure types were analyzed under SEM. Vickers indentation hardness, Young's modulus, and finite element analysis (FEA) were performed complementary to MSB tests to determine stress accumulation areas. Results: MSB results were significantly affected by the surface conditioning methods (p = 0.0001), whereas the repair composite types did not show a significant effect (p = 0.108). The interaction terms between the repair composite and surface conditioning method were also statistically significant (p = 0.0001). The lowest MSB values (MPa ± SD) were obtained in the control group (V = 4 ± 0.8; F = 3.9 ± 0.7; T = 4.1 ± 0.7) (p 0.05). While the group treated with T composite resulted in significantly lower MSB values for the HF group (T= 4.1 ± 0.8) compared to those of other composites (V = 8.1 ± 2.6; F = 7.6 ± 2.2) (p 0.05), there were no significant differences when TS was used as a conditioning method (V = 5 ± 1.7; F = 4.7 ± 1; T = 6.2 ± 0.8) (p > 0.05). The control group presented exclusively adhesive failures. Cohesive failures in composite followed by mixed failure types were more common in HF and TS conditioned groups. Elasticity modulus of the composites were 22.9, 12.09, and 10.41 GPa for F, T, and V, respectively. Vickers hardness of the composites were 223, 232, and 375 HV for V, T, and F, respectively. Von Mises stresses in the FEA analysis for the V and T composites spread over a large area due to the low elastic modulus of the composite, whereas the F composite material accumulated more stresses at the bonded interface. Conclusion: Press-on-metal ceramic could best be repaired using tribochemical silica coating followed by silanization, regardless of the repair composite type in combination with their corresponding adhesive resins, providing that no cohesive ceramic failure was observed. Schlagwörter: ceramic repair, finite element analysis, hydrofluoric acid, microshear bond strength, press-on-metal ceramic, surface conditioning method, tribochemical silica coating