Purpose: The purpose of this article was to investigate the role of control beliefs in mediating the relationship between driving cessation and change in depressive symptoms in a population-based sample of older adults. Design and Methods: We report results from a prospective, community-based cohort study that included two waves of data collected in 1992 and 1994. Participants consisted of 700 men and women aged 70 and older, including 647 drivers and 53 participants who ceased driving between baseline (1992) and follow-up (1994). Participants took part in interviews that included assessments of driving status, sociodemographic characteristics, self-rated health, sensory function, depressive symptoms (through the Center for Epidemiologic Studies-Depression scale), and expectancy of control. Using multilevel general linear models, we examined the extent to which driving status, expectancy of control, and relevant covariates explained change in depressive-symptom scores between baseline and follow-up. Results: Driving cessation was associated with an increase in depressive symptoms from baseline to follow-up. The higher depressive-symptom scores of ceased drivers relative to those of individuals who remained drivers at both waves was partly explained by a corresponding decrease in the sense of control among ceased drivers, and increased control beliefs among drivers. Implications: Interventions aimed at promoting the maintenance of personal agency and associated control beliefs could be protective against the negative psychological concomitants of driving cessation.