Reasonable prediction of freight volume is the basis for the railway department to formulate the train operation plan and organization management. In such extreme events as sudden outbreak, accurate prediction of railway freight volume data and change trend has a positive reference significance for the development of railway work. The ARIMA model was established to predict the railway freight volume from March to October 2020 by using the normal data of railway freight volume from January 2010 to January 2020 and the abnormal data from February 2020 after the outbreak of the epidemic. The results show that the ARIMA model can accurately predict the railway freight volume after extreme events by combining with abnormal data.