This paper proposes a geographic location-based distributed routing (GDR) system. GDR system is a structured distributed information lookup and routing (SDILR) system that provides information lookup, based on the geographic latitude and longitude coordinates, where each node is given the coordinates as an identifier (ID) and manages an overlay routing table. We generate an ID to reflect the geographical location without using Space Filling Curve (SFC). By transforming the 2-D plane into a grid with several uniform size non-overlapping rectangular areas. The ID is in cartesian format (x; y), which represents the longitude and latitude coordinates, respectively. In a system with size N, each node has a routing table of size logN and a search is possible in O(logN). We evaluate the performance on the mean of query path and relay distance, which is the total distance on query paths, of GDR, CAN, Chord, and Kademlia. Our results show that the mean path length of GDR, Chord, and Kademlia are the same, while the mean relay distance of GDR is shorter than that of Chord and Kademlia. Furthermore, while the mean relay distances of GDR and CAN are the same, but the mean path length of GDR is shorter than that of CAN.