This technical note presents a behavior-based navigation method for a mobile service robot with a three-wheeled omnidirectional mobile platform. A heuristic fuzzy Kohonen clustering network (FKCN) is presented to obtain laser area weights and desired robot motion heading for all types of indoor environments. An efficient behavior-based navigation structure is then proposed to merge the heuristic FKCN with three fuzzy behaviors in order to navigate the robot without any collisions in crowded and cluttered indoor environments. Simulation and experimental results are conducted to verify the feasibility and effectiveness of the proposed method.