Abstract: This paper designed a railway land management system to strengthen the supervision and comprehensive utilization of railway land, and proposed a spatial data fusion algorithm based on scale invariant feature transformation. Through the steps of keypoint extraction, feature point recognition, keypoint position and scale determination, direction extraction, and association relationship establishment, the paper concatenated and fused the maps from different railway group companies to form a railway land database, which implemented functions such as ownership management, asset disposal, and comprehensive development. It plays a positive role in promoting the goal of national railway land data linkage, overall planning, and unified management.