Abstract: A planning and navigation algorithm was proposed for intelligent wheelchair in dynamic environments. Having a hierarchical structure, this algorithm adopted a selflocaliztion method based on mapmatching, combined preplanning with online replanning for getting a semioptimal path in dynamic environments, and designed a behaviorbased control and behaviorselection method based on the local scan map. The experimental results of the intelligent wheelchair prototype show that the proposed algorithm can get an optimized path and improve safety in dynamic environments.