Aiming at the problem of how the collaborative robot can accurately grasp the specific target workpiece and avoid obstacles autonomously in the humancomputer integrated environment in the complex grasping scene of multiple objects, a visuallyguided grasping method based on supervised learning is proposed.Firstly, the supervised learning algorithm is used to train and process the image information of the object to be captured to realize the rapid recognition and location of the feature of the captured object.Then the positioning information of the grab is transmitted to the ROS system control robot through the serial port communication protocol to complete the motion planning and obstacle avoidance motion for accurate grabbing.Through many experiments and tests, it is shown that the method based on supervised learning can accurately locate the object and the cooperative robot can accurately grasp the object.This method has certain practical application value.