Objective To develop a fall risk assessment scale for older inpatients, and to promote assessment accuracy.Methods We integrated the Johns Hopkins Fall Risk Assessment Tool and Hendrich Ⅱ Fall Risk Model.Then a prospective observational study was performed among 42 490 older inpatients. Incidence of falls and the related factors were collected.Potential risk factors were identified through chi-square test and logistic regression analysis for the development of a combined risk assessment scale.At last, the scale was validated in 1 663 older inpatients.Results Logistic regression analysis identified 18 risk factors, then value of each risk factor was assigned based on the odds ratio, with a total score of 100.Validation results indicated that the area under the ROC curve for the combined fall risk assessment scale was 0.825 (95%CI:0.798-0.852).The optimal cutoff value was 7.5 points, with a sensitivity of 0.736 and specificity of 0.771.Conclusion The combined fall risk assessment scale for older inpatients can effectively predict fall risk of elderly inpatients, and help nurses to identify the high-risk patient, so as to provide targeted interventions.