This study evaluated the effect of metal surface conditioning, application of a silicon layer, water storage, and resin filling on tensile bond strength of a metal-resin system using three experimental composites (un-, micro-, and macrofilled) having the same self-curing resin composed of Bis-GMA and TEGDMA (2:1 wt%). Test specimens were prepared by bonding the resin between pairs of Ni-Cr-Be alloy cast disks (diameter, 8 mm) previously subjected to heat treatments simulating porcelain firing procedures. A specially constructed apparatus facilitated the absolutely parallel alignment and orientation of the disk faces to each other, maintaining a constant resin thickness of 100 microns. Before being bonded, the sand-blasted metal surfaces were either electrolytically etched and/or silicoated. Prior to being tested, assemblies were stored in water at 37 degrees C for one and 30 days. Thereafter, the specimens were processed in a universal testing machine at a cross-head speed of 2 mm/min until failure. Bond strengths ranged from 4.2 to 20.5 MPa. Data were analyzed by ANOVA with a factorial design (conf. level = 99%). The results showed that: (i) bond strength was increased when the metal was silicoated, (ii) the combination of sandblasting/silicoating produced the best values, and (iii) the 30-day water storage combined with silicoating enhanced the strength of the bond. The resin filling had no significant effect, indicating that neither its presence nor type affects bonding strengths to metal.