We discuss mobile robot navigation under linguistic instructions of a route. The instructions given to the robot describe the route using landmarks such as intersections, for example, “turn right at the 2nd intersection, and ….”However, it is difficult to give the robot landmark models so that it recognizes each landmark properly as a human does. We have already proposed an algorithm for mobile robot navigation, which verifies the result of landmark recognition considering the positional relationship among landmarks. In this paper, we describe the motion planning to acquire additional information effectively for navigation of the mobile robot equipped with a range sensor based on the algorithm, aiming at realizing the real robot navigation using the linguistic instructions. The environment is described by a graph based on skeletons of its free space. The motion of the robot is first roughly planned with the graph-based global map and then precisely planned with a geometric local sensory information. In experimental results using a real robot, the robot reached the goal specified by the instructions despite of errors in sensing and motion.