It is well known that the apposition of bone at implant surface would be influenced by the microstructure of titanium implants. The purpose of this study was to compare bone healing around the screw-shaped titanium implant with three different surface topographies in the canine mandibles by histological and biomechanical evaluation. All mandibular premolars of six mongrel dogs were extracted and implants were placed one month later. The pure titanium implants had different surface topographies: smooth and machined ($Steri-OSS^{(R)}$: Group II); sandblasted and acid-etched ($ITI^{(R)}$, SLA: Group III) surface. The fluorescent dyes were injected on the 2nd (calcein), 4th (oxytetracycline HCI) and 12th (alizarin red) weeks of healing. Dogs were sacrificed at 4 and 12 weeks after implantation. The decalcified and undecalcified specimens were prepared for histological and histo-metrical evaluation of implant-bone contact. Some specimens at 12 weeks after implantation were used for removal torque testing. Histologically, direct bone apposition to implant surface was found in all of the treated groups. More mature and dense bone was observed at the implant-bone interface at 12 weeks than that at 4 weeks after implantation. Under the fluorescent microscope, thick regular green fluorescent lines which mean early bone apposition were observed at the implant-bone interface in Group III, while yellow and red fluorescent areas were found at the implant-bone interface in Group I and II. The average implant-bone contact ratios at 4 weeks of healing were 54.3% in Group I, 57.7% in Group II and 66.2% in Group III. In Group I, implant-bone contact ratio was significantly lower than Group II and III(p<0.05). The average implant-to-bone contact ratios at 12 weeks after implantation were 64.3% in Group I, 66.7% in Group II and 71.2% in Group III. There was no significant difference among the three groups. In Group I and II, the implant-bone contact ratio at 12 weeks increased significantly in comparison to ratio at 4 weeks(p<0.05). The removal torque values at 12 weeks after implantation were 90.9 Ncm in Group I, 81.6 Ncm in Group II and 77.1 Ncm in Group III, which were significantly different(p<0.05). These results suggest that bone healing begin earlier and be better around the surface-treated implants compared to the smooth surface implants. The sandblasted and acid-etched implants showed the most favorable bone response among the three groups during the early healing stage and could reduce the waiting period prior to implant loading.