The aim of this study was to evaluate the influence of different cleaning regimens on the microshear bond strength (μSBS) of three different all-ceramic surfaces after saliva contamination. Cubic ceramic specimens (3 × 3 × 3 mm(3) ) were prepared from three types of ceramics: zirconium dioxide (Z), leucite-reinforced glass ceramic (E), lithium disilicate glass ceramic (EX; n = 12/subgroup). A total of 144 composite resin cylinders (diameter: 1 mm, height: 3 mm) were prepared. Three human-saliva-contaminated surfaces of ceramic specimens were cleaned with either water spray (WS), with 0.5% sodium hypochlorite solution (HC), or with a cleaning paste (CP). Control surface (C) was not contaminated or cleaned. Composite cylinders were bonded to each surface with a resin luting cement. All specimens were stored at 37°C in deionized water until fracture testing. μSBS tests were performed in a universal testing machine (0.5 mm/min), and the results (MPa ± SD) were statistically analyzed (two-way ANOVA, Bonferroni a = 0.05). Fractured surfaces were analyzed to identify the failure types using an optical microscope at 50× magnification. Two representative specimens from all groups were examined with scanning electron microscopy. μSBS test results were significantly affected by the saliva cleaning regimens (p = 0.01) and the ceramic types (p = 0.03). The interaction terms between the ceramic type and saliva cleaning regimen were also significant (p < 0.05). There were no significant differences among the μSBS values (MPa ± SD) for the Z group (C = 17.5 ± 8.8; WS = 16.0 ± 4.9; HC = 17.6 ± 5.8; CP = 16.6 ± 7.5; p > 0.05). In the EX group, C resulted in significantly higher μSBS values (32.6 ± 7.4) than CP (17.4 ± 8.9), WS (15.6 ± 7.3), and HC (14.3 ± 4.5) (p < 0.05); however, C (20.4 ± 7.1) and HC (19.2 ± 7.5) showed higher μSBS values than CP (13.8 ± 4.8) and WS (10.9 ± 5.7) in the E group. Some cohesive failures within the luting resin were observed in the E and EX groups, whereas only adhesive failures were seen in zirconia groups for all surface treatments. Different ceramic surface cleaning regimens after saliva contamination of the zirconium dioxide revealed μSBS similar to the control group, whereas all surface cleaning regimens tested significantly decreased the bond strength values in the lithium disilicate glass ceramic. The leucite-reinforced glass-ceramic group benefited from 0.5% sodium hypochlorite solution cleaning with increased bond strengths. Adhesive cementation of zirconia presents a clinically challenging protocol, and the cementation surface contamination of the zirconia restorations and the inadequate removal of the contaminants increase the risk of failure, as for all ceramic types. This study demonstrated that surface cleaning regimens should be applied according to different ceramic properties.