The study evaluated the influence of the number of retention grooves and the retainer wing thickness on the retention of resin-bonded attachments (RBAs). Overall, 64 extracted human teeth were prepared, whereby the number of retention grooves (4, 2, 1, or no grooves) and the material thickness (0.7 or 0.5mm) of the RBAs were varied. This resulted in eight groups with eight specimens each. A sample size of eight specimens per group was chosen because the chewing simulator used for dynamic loading can load eight specimens at a time. After the size of the adhesive area was determined, the abutment teeth were provided with RBAs and subjected to a dynamic loading with thermal cycling. Finally, the failure load was examined using a retention test. The mean failure loads ranged from 152 ± 50N to 228 ± 32N. There had been a significant interaction between the two main factors, i.e., number of retention grooves and material thickness. Neither the number of retention grooves nor the size of the adhesive area had a significant influence on the failure load. Using a reduced number of retention grooves seems reasonable. However, clinical considerations suggest using two retention grooves to simplify the preparation and bonding procedures. The retention of resin-bonded attachments is promising and supports the clinical application of this minimally invasive treatment modality.