Background: Annual foot checks are recommended in patients with diabetes mellitus (DM) to identify those at risk of foot ulceration. Systematic reviews have found few studies evaluating the predictive value of tests in community-based diabetic populations. Aim: To quantify the predictive value of clinical risk factors in relation to foot ulceration in a community population. Methods: A cohort of 1192 people with diabetes receiving care in community settings was recruited and a screening procedure, covering symptoms, signs and diagnostic tests was conducted at baseline. At an average 1-year follow-up patients who developed a foot ulcer were identified by an independent blind assessor. Multivariable analysis was performed to identify clinical predictors of foot ulceration. Findings: The incidence of foot ulceration was 1.93% [95% confidence interval (CI) 1.27-2.89). Three time-independent clinical predictors with five factors were selected: previous amputation [ odds ratio (OR) 14.7, 95% CI 3.1-69.5), use of insulin before 3 months with inability to distinguish between cool and cold temperatures (OR 2.97, 95% CI 1.9-4.5) and failure to obtain at least one blood pressure reading for the calculation of ankle-brachial index with the failure to feel touch with a 10-g monofilament (OR 1.7, 95% CI 1.3-2.2). Interpretation: Recommendations for annual diabetic foot check in low-risk, community-based patients should be reviewed as absolute events of ulceration are low. The accuracy of foot risk assessment tools to predict ulceration requires evaluation in randomized controlled trials with concurrent economic evaluations.