The objective of this study was to evaluate the effect of the hydrofluoridric acid (HF) concentration and time of acid conditioning on bond strength of three glass ceramics to a resin cement. Thus, fifty blocks (10 mm x 5 mm x 2 mm) of each ceramic (LDCAD: IPS e.max CAD; LCAD: IPS Empress CAD and LDHP: IPS e.max Press) were made and embedded in acrylic resin. The surfaces were polished with sandpaper (#600, 800, 1000, and 1200 grits) and blocks were randomly divided into 15 groups (n = 10) according to the following factors: Concentration of HF (10% and 5%), conditioning time (20 s and 60 s) and ceramic (LDCAD, LDHP, and L). After conditioning, silane (Prosil / FGM) was applied and after 2 min, cylinders (Ø = 2 mm; h = 2 mm) of dual resin cement (AllCem / FGM) were made in the center of each block using a Teflon strip as matrix and light cured for 40 s (1,200 mW/cm2). Then, the samples were thermocycled (10,000 cycles, 5/55°C, 30s) and submitted to the shear bond test (50 KgF, 0.5 mm/min). The data (MPa) were analyzed with 3-way ANOVA and Tukey's test (5%). Failure analysis was performed using a stereomicroscope (20x) and a scanning electron microscope (SEM). ANOVA revealed that the "concentration" factor (p = 0.01) and the interaction "acid concentration X ceramic" (p = 0.009) had a significant effect, however, the "ceramic" (p = 0.897) and "conditioning time" (p = 0.260) factors did not influence the results. The LDHP10%60s (10.98 MPa)aA* group presented significantly higher bond strength than LDHP10%20s (6.57 MPa)bA, LCAD5%20s (6,90 ±3,5)aB and LDHP5%60s (5.66 ± 2,9MPa)aA* groups (Tukey). Failure analysis revealed that 100% of specimens had mixed failure. In conclusion, etching with 5% HF for 20 seconds is recommended for lithium disilicate and leucite-reinforced CAD/CAM ceramics. However, for pressed lithium disilicate ceramic, 10% HF for 60 s showed significantly higher bond strength to resin cement.