The objective of this in vitro study was to evaluate the microleakage in ceramic inlays using different resin cements with margins in enamel and cementum/dentin interfaces. Standard Class II MOD inlay cavities were prepared in 32 noncarious human premolars. The cavities were randomly divided into 4 groups (n = 8): Control group: cavities were treated with Single Bond and incrementally filled with a composite resin (P60); Enforce group: feldspathic ceramic inlays were luted using Prime & Bond 2.1 and Enforce; RelyX group: inlays were cemented with Single Bond and RelyX ARC; Resin Cement group: ceramic inlays were bonded using Single Bond and Resin Cement. Ceramic inlays were previously treated with 10% hydrofluoric acid for 2 min, followed by silane application. After 7 days of storage in distilled water, teeth were submitted to thermocycling. After applying nail varnish, specimens were immersed in 2% aqueous solution of methylene blue for 8 h. After washing, teeth were cut into three sections through the restorations, and the leakage was assessed using a standardized score. Data were submitted to statistical analysis using nonparametric tests (Mann-Whitney and Kruskal-Wallis). Dye leakage at margins in enamel was statistically lower (p < 0.01) than at cementum/dentin interfaces. RelyX ARC performed better (p < 0.05) than resin cement (enamel) and composite restorations (cementum/dentin). No other statistical differences were observed. Both the material and the substrate interface influenced microleakage of the ceramic inlays.