Gait speed worsens with the presence of obesity, and is a powerful marker of functional dependence. Accordingly, gait speed could be a factor that improves or worsens the relationship between obesity and dependence in activities of daily living (ADL). However, to date this potential role has not been examined and the minimum gait speed threshold in the relationship between obesity and ADL is not known. The aim of this study was to determine whether speed moderates the association between obesity and dependence in ADL, and also define the gait speed threshold of this relationship. A total of 20,507 community-dwelling older adults from a cross-sectional analysis of national survey data - the Colombian Health, Well-being and Aging study (SABE, 2015) - were surveyed. The research data were collected using structured questionnaires, including basic information, ADL measured using the Barthel Index, body mass index, and gait speed (3 m). The Johnson-Neyman technique was applied to determine the gait speed threshold adjusted for age, sex and comorbidities. Regression analysis showed a significant detrimental effect of obesity on dependence in ADL, which was moderated by gait speed (β = 0.081; 95%CI: 0.045 to 0.117; p < 0.001). Adjusted for major covariates, the Johnson-Neyman technique defined two gait speed thresholds: < 0.77 m/s, indicating an aggravated adverse effect; and >1.06 m/s, indicating a positive effect. The adverse effect of obesity on dependence in ADL is moderated by gait speed. Considering these thresholds, the distribution of older adults in each of the proposed areas of significance were: below 0.77 (m/s) = 14,324 (70.0%), above 1.06 (m/s) = 1553 older adults (7.5%) and between areas = 4630 older adults (22.5%).