Abstract: This research has successfully developed an intelligent chess-playing robot based on deep Q-network, which has achieved innovation in machine vision, game algorithms, and mechanical structure. Based on yolov5 target detection technology and binocular stereo matching technology, the system realizes high-precision recognition of chess pieces (accuracy 99%) and three-dimensional spatial positioning (error ± 1.2 mm). By combining the improved minimax algorithm and alpha-beta pruning technology, the robot shows high efficiency in game tree search and strategy evaluation. In addition, using a multi-degree freedom articulated manipulator ensures high performance in chess. This research not only shows the advancement of interdisciplinary integration in technology but also shows the efficient and reliable performance in practical application, which provides a new idea for the automation and intelligent development of chess games. The experimental results show that the chess robot has high stability and accurate recognition ability to ensure the accuracy of each drop. At the same time, it has been tested in the game with human chess players, showing excellent performance.