Background: Peripheral artery disease (PAD) is associated with impaired quality of life and significant cardiovascular morbidity and mortality, yet remains under recognized and under diagnosed. Objectives: This study sought to develop and validate a risk model for the identification of individuals at risk for PAD that could be useful in the clinical setting. Methods: Twenty three variables assessed in ≈3.2 million self-referred participants without established cardiovascular disease from 2003 to 2008 who completed a medical and lifestyle questionnaire in the United States were evaluated by screening ankle brachial indices <0.90 for PAD. Subjects were divided into a derivation cohort (1.57 million) and a validation cohort (1.57 million). Lasso variable selection was used in the derivation cohort to develop the best-fitting parsimonious prediction models. Discrimination and calibrations was evaluated using the C statistic and the Hosmer-Lemeshow calibration statistic. Results: The overall prevalence of PAD was 3.96%. Using lasso variable selection, 11 variables were included in complex best-fitting model: age, sex, race, marital status, BMI group, smoking status, hypertension, diabetes, family history of PAD, physical activity, and inter-arm systolic blood pressure difference. In the validation cohort, the C-statistics for this model was .746 and the calibration was excellent (P=0.38; no significant deviation between predicted and observed outcomes). The current risk score has improved discrimination and calibration compared with other existing risk scores for PAD. Conclusion: This robust PAD risk calculator derived from a diverse population across the US provides a good risk estimate of PAD and is anticipated to assist in identifying subjects at risk for PAD.