Objective The study estabfished a model to pre-dict the weekly incidence of bacillary dysentery in Chaoyang District,and evaluated its predictive effects. Methods To eliminate the factors of sea-son-changing by means of Time Series. Auto regressive integrated moving average(ARIMA), based on model identification, estimation andverifica-tion of parameter, and analysis of the fitting of model, was established. Fi-nally,the predictive model was established by the multiple of ARLMA and seasonal factors. Results The error of the model for the prediction was -0.06 on average. The relative error was 2.32% on average. Conclusion Time series could not only accurately predict useing the data which was collected every week,but shorten the cycle of prediction.