Objective The main objective was to construct a risk prediction model of functional impairment in the elderly. Methods A total of 481 elderly patients who were admitted to four tertiary hospitals, including Shanghai Ruijin Hospital, from December in 2019 to December in 2020 were enrolled. The basic information, comorbidities, geriatric syndrome, and test results were collected. The functional impairment in 231 patients were assessed. LASSO and Logistic regression were adopted to screen predictors and construct a Nomogram. The calibration curve, decision analysis curve and concordance index(C index) were used for evaluation. Results Nine predictors including age group, gender, social activity, stroke, diabetes, anxiety, cognitive impairment, frailty and albumin were screened to construct a nomogram of impaired function in the elderly, and the calibration curve and decision analysis curve were used to validate the model, which indicated good predictive performance and clinical benefit. C index=0.938 [95% confidence interval (CI)=0.917-0.959] showed good discrimination. Conclusions The Nomogram established in the study has shown good predictive ability, clinical benefit and differentiation, which can be used easily and quickly to assess the risk of functional impairment in the elderly and provide an intuitive and effective guidance tool for clinical prediction of functional impairment in the elderly.