Railway freight volume is limited by the influence of social construction, investment and policy regulation. In order to provide better deployment support for the freight system, based on the influence factors of railway freight volume and strengthening BP prediction performance, the gray genetic BP is proposed Neural network railway freight volume prediction model (GR_GA_BP). The grey correlation analysis of railway freight volume is used to find out the index with higher correlation with railway freight volume, which is used as GR_ GA_ The input of BP prediction model predicts the freight volume.By comparing the GR_GA_BP model with other models through MAPE, RMSE and MAE, the experimental results show that the GR_GA_BP model is superior to other models and can be used as a new method to predict and study railway freight capacity.