Intraoral repairs of ceramic fixed-dental-prostheses (FDP) often include cervical recessions that require pretreatment of the exposed tooth surfaces either before or after the ceramic is conditioned with hydrofluoric (HF) acid gel. The sequence of repair protocol may cross-contaminate the exposed etched enamel or dentin surfaces during the application or rinsing process and thereby affect the adhesion. This study evaluated the influence of HF acid gel with two concentrations on bond strengths of composite to enamel and dentin. Human third molars (N=100, n=10 per group) with similar sizes were selected and randomly divided into 10 groups. Flat surfaces of enamel and dentin were created by wet ground finishing. Before or after the enamel (E) or dentin (D) was conditioned with phosphoric acid (P), substrate surfaces were conditioned with either 9.5% HF (HF(9.5)) or 5% HF (HF(5)). Subsequently, a bonding agent (B) was applied. The experimental groups by conditioning sequence were as follows where the first letter of the group abbreviation represents the substrate (E or D) followed by the acid type and concentration: group 1 (EPHF(9.5)), group 2 (EPHF(5)), group 3 (EHF(9.5)P), group 4 (EHF(5)P), group 5 (DPHF(9.5)), group 6 (DPHF(5)), group 7 (DHF(9.5)P), and group 8 (DHF(5)P). Group 9 (EPB) and group 10 (DPB) acted as the control groups. Repair resin was adhered incrementally onto the conditioned enamel and dentin in polyethylene molds. Each layer was photo-polymerized for 40 seconds. All specimens were thermocycled (×1000, 5°-55°C) and subjected to shear test (universal testing machine, 1 mm/min). Specimens that debonded during thermocycling were considered as 0 MPa. The bond strength data were analyzed using Kruskal-Wallis test and failure types using the chi-square test (α=0.05). Overall, the bond results (MPa) were lower on dentin than on enamel (p<0.01). EPB (25.6 ± 6.6) and DPB (20.2 ± 4.9) control groups showed significantly higher results than those of other groups (p<0.05). While higher mean bond strengths were obtained in group 1 (EPHF(9.5)) (11.5 ± 2.1) and group 2 (EPHF(5)) (7.3 ± 0.6), lower results were obtained when HF acid gels were applied prior to phosphoric acid (EHF(9.5)P: 5.0 ± 1.1, EHF(5)P: 3.6 ± 0.1) (p<0.05). On dentin, the results were the lowest in group 8 (DHF(5)P: 1.5 ± 1.6), being significantly lower than those of group 5 (DPHF(9.5)) (p<0.05). Scanning electron microscope (SEM) images revealed predominantly mixed failures with less than half of the composite left on both enamel and dentin surfaces (64 out of 80) (p<0.05), indicating that in general, adhesion was not ideal. Contamination of the enamel or dentin surfaces with HF acid gel impairs the bond strength of composites. Considering both the bond strength results and failure types, when dental tissues are to be repaired next to ceramic, application of phosphoric acid before HF acid gel application can be recommended. HF acid gel concentration did not influence the results except on enamel.