The navigation of mobile robots is a vital aspect of technology in robotics. We applied the D++ algorithm, which is a novel and improved path-planning algorithm, to the navigation of mobile robots. The D++ algorithm combines Dijkstra's algorithm with the idea of a sensor-based method, such that Dijkstra's algorithm is adapted to local search, and the robot can determine its next move in real-time. Although the D++ algorithm frequently runs local search with limited ranges, it can compute optimum paths by expanding the size of the searching range to avoid local minima. In addition, we verified the performance of the D++ algorithm by applying it to a real robot in a number of environments. The use of the D++ algorithm enables robots to navigate efficiently in unknown, large, complex and dynamic environments.