The reliability of adhesive bonding to CAD/CAM resin composites is influenced not only by material composition and surface treatment but also by the testing methodology used to assess bond strength. However, the impact of different shear bond strength (SBS) test configurations remains insufficiently clarified. This study evaluated the influence of different surface pretreatment protocols and SBS test methods on the bonding performance of a self-adhesive resin cement to two CAD/CAM materials: a conventional particulate-filled composite (Cerasmart 270) and an experimental short glass fiber-reinforced composite (SFRC CAD). Specimens (14 × 12 × 3 mm; n = 80 per material) were ground with 320-grit silicon carbide paper and divided according to surface pretreatment: airborne-particle abrasion (APA) or APA followed by hydrofluoric acid application for 60 s (APA + HF). Each group was further subdivided based on the SBS test method using either resin cement cylinders fabricated with a custom transparent mold (diameter: 3.6 mm; height: 3 mm) or metallic cylinders cemented to the treated surface. Half of the specimens were tested after 48 h of water storage, while the remainder underwent hydrothermal aging by boiling in water for 16 h prior to testing. Material type, SBS test method, surface pretreatment, and aging significantly affected bond strength (p < 0.05). The metallic cylinder method produced higher SBS values than the transparent mold technique, particularly for SFRC CAD. APA + HF tended to reduce SBS in Cerasmart 270, particularly after aging, whereas SFRC CAD showed comparable or higher bond strength values with APA alone. Aging decreased SBS in most groups. Overall, bond strength was influenced by both material type and test methodology. Within the limitations of this study, airborne-particle abrasion alone may be sufficient for SFRC CAD materials, while additional HF treatment may not provide further benefit. These findings highlight the importance of considering both material characteristics and test configuration when interpreting laboratory bond strength data.