Purpose: The purpose of this study was to investigate the effect of sire and shape of retention element on the bond strength of indirect composite resin and metal. Material and method: The metal disk specimens, each 6mm in diameter, were cast from CrCo alloy. They were divided into 8 groups by applied retention element. retention bead group $B2;({phi};0.2mm),;B4;({phi};0.4mm),;B6;({phi};0.6mm),;B8;({phi};0.8mm)$, retention crystal group C2 (0.2mm), C5 (0.5mm), C8 (0.8mm) and sandblasting group SB ($110{mu}m;Al_2O_3$ blasting) as control. Eighty-eight metal specimens were veneered with $TESCERA^{(R)}$ Indirect resin system. One specimen of each group was sectioned and the resin-metal bonding pattern at the interface was observed under measuring microscope. Other specimens were then tested for tensile bond strength on an Instron universal testing machine at a crosshead speed of 2mm/min. Results: 1. Compared to sandblasting, beads or crystals increased the resin-metal bond strength (P<.05). 2. 0.2mm retention crystals were most effective in improving the resin-metal bond strength (P>.05). 3. 0.2mm beads showed the highest bond strength among retention bead groups, but there was no statistically significant difference (P>.05). 4. Retention crystals tend to be higher in bond strength than retention beads due to wider surface area. 5. The larger retention element, the larger the undercut for the mechanical retention, but the gap at resin-metal interface was also increased. Conclusion: Within the limitations of this study, 0.2mm retention crystals were most effective in improving the resin-metal bond strength.