Abstract: This paper designs a machine vision-based robotic arm gripping system based on hardware such as a six-degree-of-freedom Kinova robotic arm and a Realsense depth camera. To solve the problem that the traditional template matching method was not rotation invariant, the color recognition algorithm was applied to the target recognition. A parametric model of the robot system was established, the image was pre-processed, and the center-of-mass coordinates of the target were obtained by the color recognition algorithm. The kinematic model of the robot arm was established by the D-H parameter method, and the joint angle was solved by the inverse kinematics to complete the grasping. The experimental data show that the system designed in this paper is highly feasible.