Objective Frailty and pre‐frailty are two common conditions in the older people, but whether these conditions could predict depression is still limited to a few longitudinal studies. In this paper, we aimed to investigate whether frailty and pre‐frailty are associated with an increased risk of depression in a prospective cohort of community‐dwelling older people. Methods Four thousand seventy‐seven community‐dwelling men and women over 60 years without depression at baseline were included from the English Longitudinal Study of Ageing. Frailty status was defined according to modified Fried's criteria (weakness, weight loss, slow gait speed, low physical activity and exhaustion) and categorized as frailty (≥3 criteria), pre‐frailty (1–2 criteria) or robustness (0 criterion). Depression was diagnosed as ≥4 out of 8 points of Center for Epidemiologic Studies Depression Scale, after 2 years of follow‐up. Results Over a 2‐year follow‐up, 360 individuals developed depression. In a logistic regression analysis, adjusted for 18 potential baseline confounders, pre‐frailty (odds ratio (OR) = 0.89; 95% confidence interval (CI), 0.54–1.46; p = 0.64) and frailty (OR = 1.22; 95% CI, 0.90–1.64; p = 0.21) did not predict the onset of depression at follow‐up. Among the criteria included in the frailty definition, only slow gait speed (OR = 1.82; 95% CI, 1.00–3.32; p = 0.05) appeared to predict a higher risk of depression. Conclusions Among older community dwellers, frailty and pre‐frailty did not predict the onset of depression during 2 years of follow‐up, when accounting for potential confounders, whilst slow gait speed considered alone may predict depression in the older people. Copyright © 2017 John Wiley & Sons, Ltd.