OBJECTIVE To compare the cusp fracture resistance of teeth restored with composite resins and two post systems. METHODS Eighty extracted single-rooted human maxillary premolars were randomly assigned to eight groups (n = 10). Group 1 (control) did not receive any preparation. From groups 2 to 8, the teeth were root filled and mesio-occluso-distal (MOD) cavities were prepared. Group 2 remained unrestored. Group 3 was restored with packable resin composite using a single-step adhesive. Group 4 was restored with packable resin composite using a single-step adhesive and a thin layer of flowable resin composite. Group 5 was restored with packable resin composite using a total-etch two-step adhesive. Group 6 was restored with ormocer resin composite using a total-etch two-step adhesive. Group 7 was restored with an endodontic glass fibre post and hybrid resin composite using a total-etch two-step adhesive. Group 8 was restored with an endodontic zirconium post and hybrid resin composite using a total-etch two-step adhesive. The teeth were then mounted in a universal testing machine, the buccal cusp loaded (30 degrees ) until fracture, and the data analysed statistically. RESULTS Group 1 had the greatest fracture resistance, and group 2 the poorest. Groups 5-8 had significantly greater (P < 0.05) fracture resistance than groups 3 and 4. No significant differences were found between groups 3 and 4, or amongst groups 5-8 (P > 0.05). CONCLUSIONS For root filled maxillary premolars with MOD cavities, adhesive resin composite restorations, with and without glass and zirconium posts, increased the fracture resistance of the buccal cuSPS. A total-etch two-step adhesive increased significantly fracture resistance more than a one-step adhesive. For the one-step adhesive, an additional layer of flowable resin composite did not enhance fracture resistance.