In　addressing　the　challenge　of　obstacle-avoidance　motion　planning　for　space　redundant　manipulators　operating　in　environments　with　obstacles,　an　improved　method　based　on　a　rapidly　exploring　random　tree　(RRT)　algorithm　is　proposed.　The　method　first　adopts　a　target　probabilistic　sampling　strategy　to　enhance　the　target　orientation.　Secondly,　an　artificial　potential　field　is　constructed　to　steer　the　growth　of　the　node　tree　in　joint　space.　Subsequently,　a　collision　detection　model　between　the　manipulator　and　obstacles　is　established　based　on　the　bounding　capsule　and　bounding　sphere　in　Cartesian　space.　Finally,　the　paths　are　pruned　and　smoothed　with　a　cubic　Bezier　curve.　The　simulation　results　validate　the　method’s　capacity　to　effectively　plan　a　collision-free　path　for　the　space　redundant　manipulator　and　the　joint　motion　trajectory　is　smooth.