Abstract: Objective The present study was conducted to explore the feasibility of time-series ARIMA model in the application of early warning of measles. Methods SAS 9.0 statistical software was used to conduct modeling of ARIMA model by weekly incidence of measles based on the measles cases reported through the National Management Information System for Disease Reporting Results It was found that season model ARMA(1,53) of measles case number series, together with comparison with the number of measles cases in 2005, contributed to timely and accurate warning. Conclusion ARIMA model can be applied to early warning of measles, providing a scientific basis for the prevention and control of the epidemic.