Abstract: This paper proposes a new index structure to efficiently store and retrieve the past，present and future positions of moving objects in networks，L2R-Tree.The index structure consistes of two-level 2D R-tree and a linklist.The two-level R-trees are used to index polyline and objects’ movements along the polylines.The present positions and future prediction trajectories of moving objects are stored in linklist.L2R-Tree can not only support the trajectories query of moving objects in road networks but also is especially easy to query all the moving objects in the same route by lengthways list.Moreover，point query and range query are implemented based on this index structure.Finally，experimental studies indicate that the L2R-tree index outperformed TPR-tree in index creation as well as in querying.