Based on the freight volume data of the Changjiang River main line from 2005 to 2020, the grey system GM(1,1) model, the quadratic exponential smoothing GM(1,1) model, and the trend extrapolation method are proposed to predict freight volume of the Changjiang River main line. According to those methods, the historical data are calculated by Matlab software and those results are compared and analyzed. At last, the residual periodic correction method is used to correct the predicted results of models, and the final predicted results from 2021 to 2025 are obtained by the combined weighting of previous selected models' predicted results.