Automatic　lane-changing　is　a　complex　and　common　task　for　autonomous　vehicle　control.　In　this　study,　a　hierarchical　decoupled　path　and　velocity　planning　model　for　lane　changing　is　proposed　to　enhance　driving　safety,　comfort,　and　traffic　efficiency.　First,　a　parametric　trajectory　model　is　established　based　on　the　vehicle　kinematic　model,　and　the　initial　trajectory　is　solved　quickly　by　the　Sequential　Quadratic　Programming　algorithm;　in　addition,　the　path　optimization　function　is　designed　to　optimize　the　trajectory　curvature,　and　the　distance-based　velocity　optimization　method　is　used　to　improve　the　trajectory　transverse,　longitudinal　acceleration,　and　jerk.　To　ensure　the　accuracy　of　path　tracking,　a　comprehensive　online　trajectory　optimization　function　is　proposed　according　to　tracking　error　fitting　and　vehicle　reachability　domain.　The　validation　results　demonstrate　that　the　optimized　path　transverse　velocity,　acceleration,　and　jerk　change　curve　are　smoother,　which　meets　the　safety　and　comfort　requirements　of　trajectory　planning.　Finally,　the　feasibility　of　the　proposed　trajectory　planning　is　verified　in　a　prototype　vehicle　real-world　test.