In view of the large scale, randomness and instability of railway passenger volume data, and the difficulty to ensure the accuracy of short-term passenger volume forecast, a dynamic data integration method of urban land use based on GIS geographic information technology was proposed. The individual demand factors affecting high-speed rail travel are selected, the historical data are expanded by the correction factor, and the effective structural components are extracted by the time series method. The comfort index is integrated into the original MD model, and the prediction process of the improved MD model is set up. Through the factors affecting the travel sacrifice, the improved travel sacrifice function is designed to realize the short-term passenger volume prediction of high-speed railway. The experimental results show that the prediction accuracy of the method can reach 99.5% for the short-term passenger volume of high-speed rail in different cities, which has certain application value.