To　solve　the　problem　of　high　complexity,　safety　and　reliability　of　multi-DOF　manipulator　trajectory　planning　in　three-dimensioned　space,　based　on　the　advantages　of　probability　completeness　and　lightweight　of　rapidly-exploring　random　trees　(RRT)　algorithm　in　high-dimensioned　space,　a　target　heuristic　RRT　based　on　uniform　probability　(PH-RRT)　was　proposed.　First,　the　threshold　of　probabilistic　sampling　was　selected　as　the　node　standard　based　on　the　distribution　mechanism　of　uniform　probability,　and　it　was　compared　with　the　random　sampling　value.　When　the　random　sampling　value　was　within　the　set　threshold　value　range,　the　target　point　was　determined　as　a　random　point　for　node　expansion.　When　the　random　sampling　value　was　outside　the　set　threshold　range,　random　point　was　randomly　generated,　and　the　new　node　expanded　under　the　target　heuristic　of　the　target　gravity　and　the　random　point　gravity.　Then,　based　on　the　planned　path,　the　breadth-first　search　idea　was　introduced　to　optimize　the　planned　path　to　improve　the　smoothness　of　the　path　and　reduce　the　length　of　the　path.　Experimental　results　show　that　the　proposed　method　can　well　solve　the　problem　of　traditional　method　of　RRT　inherent　blind　search,　and　reduce　the　time　of　path　planning　and　path　length　to　improve　the　efficiency　of　path　planning　of　manipulator.　©　2022,　Editorial　Department　of　Journal　of　Beijing　University　of　Technology.　All　right　reserved.