OBJECTIVES: To examine the relationship between indicators of inflammation and the incidence of mobility limitation in older persons. DESIGN: Prospective cohort study: the Health, Aging and Body Composition Study. SETTING: Pittsburgh, Pennsylvania, and Memphis, Tennessee. PARTICIPANTS: A total of 2,979 men and women, aged 70 to 79, without mobility limitation at baseline. MEASUREMENTS: Serum levels of interleukin (IL)-6, tumor necrosis factor alpha (TNFα), and C-reactive protein (CRP) and soluble cytokine receptors (IL-2sR, IL-6sR, TNFsR1, TNFsR2) were measured. Mobility limitation was assessed and defined as reporting difficulty or inability to walk one-quarter of a mile or to climb 10 steps during two consecutive semiannual assessments over 30 months. RESULTS: Of the 2,979 participants, 30.1% developed incident mobility limitation. After adjustment for confounders (demographics, prevalent conditions at baseline, body composition), the relative risk (RR) of incident mobility limitation per standard deviation (SD) increase was 1.19 (95% confidence interval (CI) = 1.10-1.28) for IL-6, 1.20 (95% CI = 1.12-1.29) for TNFα and 1.40 (95% CI = 1.18-1.68) for CRP. The association between inflammation and incident mobility limitation was especially strong for the onset of more severe mobility limitation and when the levels of multiple inflammatory markers were high. When persons with baseline or incident cardiovascular disease events or persons who were hospitalized during study follow-up were excluded, findings remained similar. In a subset (n = 499), high levels of the soluble receptors IL2sR and TNFsR1 (per SD increase: RR = 1.23 (95% CI = 1.04-1.46) and RR = 1.28 (95% CI = 1.04-1.57), respectively) were also associated with incident mobility limitation. CONCLUSION: Findings suggest that inflammation is prognostic for incident mobility limitation over 30 months, independent of cardiovascular disease events and incident severe illness.