Mobile　robot　navigation　is　one　of　the　important　domains　in　mobile　robot　technologies＇　research.　This　problem　is　divided　into　two　categories　of　basic　sub-problems:　path　planning　and　motion　planning.　In　this　paper　we　integrate　the　path　planning　and　motion　planning　of　a　robot　into　a　uniform　framework,　which　is　described　by　a　hybrid　system.　Hybrid　systems　combine　discrete　and　continuous　behavior　so　that　we　can　deal　with　the　complexity　of　the　environment　at　the　discrete　level　and　dynamics　of　a　robot　at　the　continuous　level.　At　the　discrete　level,　on　the　triangulation　of　the　state　space　of　a　robot,　a　dual　graph　is　constructed　following　a　proposed　target　attractive　principle.　Then　based　on　this　dual　graph,　we　design　an　incremental　heuristic　search　algorithm　to　deal　with　dynamic　environment　and　find　an　optimal　sequence　of　adjacent　triangles.　At　the　continuous　level,　a　motion　planning　algorithms　that　can　autonomously　generate　the　translational　and　rotational　velocities　for　the　robot　to　travel　along　the　given　sequence　of　triangles.　Simulation　results　demonstrate　the　correctness　and　effectiveness　of　the　approach.