This　paper　describes　the　importance　of　robotic　grasping　in　robotics　applications　and　emphasizes　the　complexity　and　multidisciplinary　cross-application　of　grasping　tasks.　The　authors　present　some　recent　grasping　algorithms　and　methods,　including　a　planar　grasping　model　based　on　RGB　2D　image　information,　the　application　of　depth　cameras,　and　an　end-to-end　6-DOF　grasping　bit-position　detection　network.　The　authors　also　propose　an　improved　algorithm　to　generate　reliable　grasping　schemes　through　specific　target　detection　in　depth　images　to　meet　the　challenges　of　grasping　in　complex　environments.　The　article　also　presents　the　overall　design　of　a　ROSbased　robotic　arm　grasping　system　and　experimentally　verifies　the　feasibility　and　accuracy　of　the　improved　algorithm.　The　results　show　that　the　improved　algorithm　achieves　92%　success　rate　in　multiple　grasping　tasks　with　real-time　performance.　This　research　is　of　great　significance　to　the　development　and　application　of　robot　grasping　technology.