The replacement of defective amalgam restorations leads to loss of tooth material and weakens the tooth, creating an increased risk of cusp fracture. The repair of such defects is a minimal intervention technique. The current study compared the repair bond strengths of a resin composite to amalgam and an amalgam-dentin complex after various surface conditioning methods. The specimens (N = 50) consisted of sound human canines with cylindrical preparations (diameter: 2.3 mm, depth: 3 mm) with amalgam-dentin complex (N = 30, n = 10/per group) and two groups with amalgam only (N = 20, n = 10/per group). The teeth were embedded in auto-polymerized polymethylmethacrylate (PMMA). The preparations were filled with non-Gamma 2 amalgam. The enamel was removed to expose dentin. The specimens with the amalgam-dentin complex were randomly assigned to one of the following conditioning methods: Group 1: Silicacoating amalgam, etching dentin, silane application on amalgam, primer/bonding on dentin, opaquer on amalgam, resin composite on both; Group 2: Etching dentin, silicacoating amalgam, silane application on amalgam, primer/bonding on dentin, opaquer on amalgam, resin composite on both and Group 3: Etching dentin, primer/bonding on dentin, opaquer, resin composite. The specimens with only amalgam were assigned to one of the following conditioning methods: Group 4: Silicacoating, silane application, opaquer, resin composite and Group 5: Opaquer, resin composite. For the two control groups, where no dentin was involved (Groups 4 and 5), bonding was achieved only on amalgam and Group 5 had no conditioning. The specimens were kept in water at 37 degrees C for five weeks before bond strength (MPa +/- SD) testing (Universal Testing Machine). After debonding, the failure types were analyzed. The results were significantly affected by the surface conditioning method (ANOVA). Only dentin conditioning (Group 3) showed the highest bond strength (39.9 +/- 14). The unconditioned control group (Group 5) showed the least favorable results (1.4 +/- 0.5). Multiple comparisons (Tukey-Kramer adjustment) showed that the mean values of Group 1 (34.1 +/- 11.4), 3 (39.9 +/- 14) and 4 (35.5 +/- 4) were not significantly different (p > 0.05), but between Groups 2 (22.8 +/- 6.6) and 3 (39.9 +/- 14), significant differences were observed (p = 0.0027). For reliable repair of amalgam restorations, including dentin fractures, the amalgam surface should first be silica coated, then the dentin/enamel should be etched, washed and rinsed thoroughly. Finally, the amalgam should be silanized and primer/bonding applied onto the dentin.