To evaluate the influence of different ceramic surface conditioning methods on the fatigue failure load of adhesively cemented simplified lithium-disilicate glass-ceramic restorations. Ceramic (IPS e.max CAD, Ivoclar Vivadent) (Ø = 10 mm; thickness = 1.2 mm) and epoxy resin (Ø = 10 mm; thickness = 2.3 mm) disks were produced. The ceramic bonding surfaces were treated as follows: no etching and MPS-silane primer application only (MN); etching with 10% hydrofluoric acid (HF) for 20 s followed by primer application (HF + MN); HF + universal multimode adhesive application (HF + SU); etching with a one-step etching primer (ME&P); HF + primer + conventional adhesive (HF + MN + PAB). The epoxy resin disks were etched with 10% HF for 20 s followed by a coat of bonding agent (Multilink Primer A+B). Pairs of ceramic/epoxy resin disks were cemented with composite cement (Multilink N, Ivoclar Vivadent). The mean fatigue failure load was determined by the staircase method (100,000 cycles at 20 Hz frequency; initial load = 1435 N; step size = 72 N). ME&P had the highest fatigue failure load, followed by HF etched groups, while the non-etched condition (MN group) had the lowest. All samples presented radial cracks originating from defects at the conditioned ceramic surface (interface). The simultaneous physicochemical conditioning with one-step self-etching ceramic primer promoted the best fatigue behavior results of the glass-ceramic restorations. It might indicate that this one-step conditioning reduces the number of flaws at the ceramic surface due to the slighter surface alterations than those produced by hydrofluoric acid etching, improving the fatigue behavior.