This paper proposes a reliable high-precision positioning system that integrates a navigation satellite system and a vision system in order to resolve the problems of position errors and outdoor shadowed areas, two key disadvantages of navigation satellite systems. For positioning of an object, users should receive observation data from at least four navigation satellite signals. However, in urban areas, accurate positioning is difficult due to factors such as high buildings, obstacles, and reflected waves. In order to address the above problem, a vision system is employed in this study. First, we determine an exact position value of a specific building in urban areas which provide a poor environment for navigation satellite. Then the building is identified by the vision system and its position error is corrected by using the identified building. A moving object can identify a specific building with the vision system, make observation data, and revise positioning, thereby the result can be more reliable and accurate.