In this paper, we presented a real-time 2D human gesture grading system from monocular images based on OpenPose, a library for real-time multi-person keypoint detection. After capturing 2D positions of a person's joints and skeleton wireframe of the body, the system computed the equation of motion trajectory for every joint. Similarity metric was defined as distance between motion trajectories of standard and real-time videos. A modifiable scoring formula was used for simulating the gesture grading scenario. Experimental results showed that the system worked efficiently with high real-time performance, low cost of equipment and strong robustness to the interference of noise.