Purpose: This retrospective study was set to explore the influence of local bone density (BD) on implant cumulative survival rates (ICSRs) and to assess prognostic factors associated with implant failure at sites with different BD. Materials and Methods: Between January 2005 and December 2011, 2,684 implants were placed in 1,377 patients and included in the study. Implants at sites with different BD were divided into four groups (G1 to G4) according to the Lekholm and Zarb classification, corresponding to bone types 1 to 4. ICSRs and the reasons for failure in each group were evaluated. Factors related to the local distribution of BD were also analyzed. A number of predictive variables were examined by univariate and multivariate analyses to evaluate prognostic factors and their influence on implant failure rates. Results: In total, 45 implants were lost, resulting in ICSRs for G1 to G4 of 100%, 98.18%, 96.83%, and 92.25%, respectively. The main reasons for failure in each group were failed osseointegration and occlusal overloading. Low BD was associated with advanced age (> 50 years) and the posterior maxilla. Based on multivariate analysis, diabetes mellitus and nonthreaded implants were significant factors in the high-BD group (G2), while advanced age, smoking, nonthreaded implants, and immediate loading were risk factors for the low-BD group (G3 and G4). Conclusion: BD is one of the most important factors influencing the long-term ICSR, which decreases with decreasing BD values. Accurate risk evaluation for sites with different BD before implantation will be beneficial to implant survival. Schlagwörter: bone density, dental implants, implant cumulative survival rate, prognostic factors