As the enhancement of living standard and the influx of talents,the competition of the license plate auction in Shanghai is increasingly fierce.So the short-term prediction on the average price has high practical significance.By analyzing the correlation between the drop number,bidders number,historical data and other influential factors respectively,it is found that the short-term prediction using historical data is most effective.Accordingly,based on the three exponential smoothing method and dynamic three exponential smoothing method,using the historical data of license plate monthly average prices from January to October in 2016,the prices in November and December have been predicted.Besides,by analyzing the prediction data,it is shown that the dynamic three exponential smoothing method is better than the three exponential smoothing method.Therefore,it is concluded that using the dynamic three exponential smoothing method and historical data to predict the monthly average price of the license plate can provide an effective guidance for the bidding price of license plate.