Abstract: Heavy workload of system design and poor portability will arise when Robots are used to grasp specific objects through machine vision. To resolve this problem, a seven-axis manipulator grabbing method was proposed using hand-eye combination based on robot operating system (ROS). Under the framework of ROS, target grabbing was carried out by using cooperative robot FRANKA PANDA combined with RGBD camera. The xacro robot model with RGBD camera was created and RViz was used to visualize the 3D model. MoveIt! was used for the joint simulation of RViz and Dazebo. The image information obtained by ROS was sent to the OpenCV library for image processing and object positioning, and then the object coordinate information was sent back to ROS for robotic arm grasping task. The experiment results show that the proposed method is simple and effective, and the gripping method designed based on ROS is highly portable, which solves the problem of large-scale control changes in robot systems, especially with complex and diverse objects.