Objectives: This study examined the influence of depressive symptoms on the prevalence of physical disability in a cohort study of Beijing elderly (1992-2000) and analyses the role of some confounding variables in this relationship. Methods: A cohort of 1828 elderly aged 55 and older who were initially free of any physical disability was followed up for 8 years. Baseline depression was assessed by the Center for Epidemiological Studies Depression Scale. Disability on activities of daily living (BADL and IADL) was measured in 1994, 1997, and 2000 respectively. Results: The cumulative percentages of persons who developed BADL disability and IADL disability during each follow-up interview was significantly greater among depressed than non-depressed individuals. Compare with the 1525 non-depressed individuals, the 303 depressed individuals had a relative risk (95% confidence interval) of 2.52 (2.02, 4.82) for BADL disability and 4.98 (2.46, 10.09) for IADL disability, respectively. Although adjustment for confounding factors reduced the risk to 2.20 (1.33, 3.62) and 4.29 (2.08, 8.86), the detrimental effect of depression on disability remained significant. The items of IADL which connected physical activity had higher impairment among depressed individuals. Conclusions: Depression in elderly persons may increase the risk for prevalent disability. It is very important for elders to prevent and reduce depression for improving the quality of their life and physical function.