The New International Land-Sea Trade Corridor (NILSTC) has significant impacts on the transportation, trade, and economy of western China and Southeast Asian countries. Previous literature has proposed the use of the data envelopment analysis (DEA) model to calculate the transportation efficiency of the NILSTC. However, few scholars have integrated DEA models with machine learning to forecast future development goals. To predict the planning goals for provincial railway transportation in the NILSTC by 2025, we utilized the EATBoosting model for calculations. By selecting relevant data for the railway transportation of the NILSTC in 2022, we initially employ the traditional DEA model to calculate its railway transportation efficiency. Subsequently, we utilized the EATBoosting model to forecast the efficiency and output targets for 2025. The results indicate that: (1) the overall efficiency predicted for 2025 is lower compared to the efficiency calculated by traditional DEA models, suggesting that each province needs to enhance efficiency in the future; (2) the predicted output in 2025 is higher than the initial output in 2022, indicating that all provinces, especially Chongqing and Sichuan, need to increase their output and strive to achieve the predicted goals. In conclusion, our research has a significant impact on decision-makers when establishing planning goals.