Abstract: Objective To evaluate the intervention effect of fall prevention using Difference-in-Difference model.Methods Based on the data of fall of the elderly in Qinghai and Shanxi Provinces in 2013 and 2015, the fall-related control variables were screened by logistic regression model, and then the propensity score method and DID model were be used to evaluate the preventive effect of fall before and after intervention.Results The incidence of falls was related to before and after the implementation of the project, whether having intervention, whether suffering from cardiovascular and cerebrovascular diseases, ethnicity and age. The net effect of fall intervention was 10.8% (t=3.78, P<0.001) after control the factors related to fall.Conclusion The DID model combined with propensity score method can accurately separate the net effect of interventions and facilitate the elucidation of the causal relationship between intervention and effect.