In this paper, we describe our positioning system based ultra-compact Laser Range Finder (LRF) for mobile robot. The system consists of reflectors as landmark and a laser range finder (LRF) named "Pos-URG". The landmarks are randomly and sparsely placed in environment, and their position and posture are known. Pos-URG has map data which consists of pose information and relation information between each landmark, which are sorted in ascending order of distance value between each other landmarks. When the Pos-URG finds multiple number of landmarks, the set of landmarks in map which correspond the detected landmarks is searched. The searching is done using not only each landmark's position and posture but also relations between them. Then, Pos-URG's position and posture is calculated with coordinate transform from sensor coordinates to map coordinates. The most important characteristic of this proposed searching method is its low computational time and it enables the method to be implemented into low-cost and low-spec CPU which is assembled into Pos-URG.