The　number　of　a　redundant　manipulator’s　DOF　is　greater　than　the　number　of　the　workspace’s　dimensions,　creating　difficulty　in　solving　its　inverse　kinematics.　In　the　workspace,　each　pose　corresponds　to　infinite　number　of　an-gles　of　the　arm.　As　a　result　this　means　that　the　traditional　inverse　kinematics　so-lution　method　is　computationally　large　and　complex.　With　the　aim　of　trajectory　planning,　an　inverse　kinematics　algorithm　based　on　the　genetic　algorithm　has　been　proposed　in　this　paper　to　generate　a　smooth　trajectory.　The　initial　pose　of　the　arm　was　taken　as　the　initial　condition,　then　the　constraint　function,　a　reduction　in　the　search　interval　and　a　smooth　reward　function　was　used,　and　finally　the　genetic　algorithm　was　implemented.　It　was　found　that　this　method　greatly　reduced　the　solution　of　the　large　angular　momentum　of　the　arm　and　tended　to　choose　a　smoother　solution.　Through　the　use　of　the　constraint　function　and　the　smooth　reward　function　as　the　optimization　index,　the　inverse　kinematic　solution　corresponding　to　the　preset　trajectory　was　able　to　be　effectively　solved.　©　2019,　Springer　Nature　Switzerland　AG.