The　inverse　kinematics　problem　plays　a　crucial　role　in　robotic　manipulator　planning,　autonomous　control,　and　object　grasping.　This　problem　can　be　solved　in　simple　environments　based　on　existing　studies.　However,　it　is　still　challenging　to　quickly　find　a　feasible　inverse　kinematic　solution　when　obstacle　avoidance　is　required.　In　this　paper,　we　present　a　nonconvex　composite　programming　method　to　solve　the　inverse　kinematics　problem　with　overhead　obstacle-avoidance　requirements.　Our　method　enables　efficient　obstacle　avoidance　by　directly　calculating　the　minimum　distance　between　the　manipulator　and　the　overhead　environment.　We　construct　end-effector　error　functions　based　on　the　Product　of　Exponentials　model　and　explicitly　provide　their　gradient　formula.　We　derive　the　minimum　distance　based　on　the　geometry　parametric　equation　and　directly　utilize　it　to　construct　the　obstacle　avoidance　function.　We　propose　an　enhanced　version　of　adaptive　moment　estimation　based　on　shorttime　gradient　information　to　improve　optimization　performance.　Finally,　we　conduct　simulations　and　experiments　in　overhead　line　environments.　Comparative　results　with　other　optimization　methods　demonstrate　that　our　proposed　method　achieves　a　high　success　rate　with　a　low　solution　time.