This study evaluated the influence of distinct substrates on the mechanical fatigue behavior of adhesively cemented simplified restorations made of glass, polycrystalline or polymer infiltrated-ceramics. CAD/CAM ceramic blocks (feldspathic – FEL; lithium disilicate – LD; yttria-stabilized zirconia – YZ; and polymer-infiltrated ceramic network – PICN) were shaped into discs (n = 15, Ø = 10 mm; thickness = 1.0 mm), mimicking a simplified monolithic restoration. After, they were adhesively cemented onto different foundation substrates (epoxy resin – ER; or Ni–Cr metal alloy – MA) of the same shape (Ø = 10 mm; thickness = 2.0 mm). The assemblies were subjected to fatigue testing using a step-stress approach (200N-2800 N; step-size of 200 N; 10,000 cycles per step; 20 Hz) upon the occurrence of a radial crack or fracture. The data was submitted to two-way ANOVA (α = 0.05) to analyze differences considering ‘ceramic material’ and ‘type of substrate’ as factors. In addition, a survival analysis (Kaplan Meier with Mantel-Cox log-rank post-hoc tests; α = 0.05) was conducted to obtain the survival probability during the steps in the fatigue test. Fractographic and finite element (FEA) analyzes were also conducted. The factors ‘ceramic material’, ‘type of substrate’ and the interaction between both were verified to be statistically significant (p < .001). All evaluated ceramics presented higher fatigue failure load (FFL), cycles for failure (CFF) and survival probabilities when cemented to the metallic alloy substrate. Among the restorative materials, YZ and LD restorations presented the best fatigue behavior when adhesively cemented onto the metallic alloy substrate, while FEL obtained the lowest FFL and CFF for both substrates. The LD, PICN and YZ restorations showed similar fatigue performance considering the epoxy resin substrate. A more rigid foundation substrate improves the fatigue performance of adhesively cemented glass, polycrystalline and polymer infiltrated-ceramic simplified restorations.