Objective: To explore the association between baseline complete edentulism and the risk of incident cardiovascular disease (CVD) in older adults, and to verify the consistency across clinical subgroups and robustness of the findings, providing evidence for early CVD risk identification and comprehensive prevention and management among older adults. Methods: Data were derived for a retrospective analysis from two longitudinal cohorts: the China Health and Retirement Longitudinal Study (CHARLS) and the Health and Retirement Study (HRS). Participants aged ≥ 60 years without prevalent CVD at baseline were enrolled. Baseline complete edentulism was defined as the exposure variable, and incident CVD during follow-up was set as the outcome variable. Kaplan-Meier analysis was performed to compare differences in CVD-free survival probability between groups. Multivariable Cox proportional hazards regression models were used to estimate the independent association between edentulism and incident CVD. In addition, an interval-censored proportional hazards model was applied as a sensitivity analysis to assess the potential impact of approximating event times on the risk estimates. Hazard ratios and 95% confidence intervals were used for description. Six-dimension subgroup (age, sex, education, etc.) analyses and interaction tests were performed to confirm the generalizability of the association, and five sensitivity analyses were conducted to validate the reliability of the results. Results: A total of 5 522 participants from the CHARLS and 7 738 participants from the HRS were included. Survival analysis revealed that the CVD-free survival probability was significantly lower in the complete edentulism group than in the non-edentulous group across both cohorts (P < 0.001). Fully adjusted regression models identified complete edentulism as an independent risk factor for new-onset CVD among older adults (CHARLS: HR = 2.64, 95% CI: 2.33-2.99, P < 0.001; HRS: HR = 2.69, 95% CI: 2.46-2.95, P < 0.001). The results from the interval-censored model were highly consistent with those from the conventional Cox model. This association persisted across all predefined six dimensions subgroups with all interaction P > 0.05. Outcomes of five sensitivity analyses were consistent with primary analyses, indicating favorable robustness. The strongest association was observed when stroke was treated as a separate endpoint (CHARLS: HR = 2.85; HRS: HR = 2.91). Conclusion Findings from both the Chinese and U.S. cohorts consistently indicate that complete edentulism independently elevates the risk of incident CVD in older adults, with robust associations across diverse subgroups. Complete edentulism may serve as an auxiliary indicator for CVD risk stratification among older adults.