The highly intelligent and fast-paced life in modern society has made the requirements of higher efficiency and higher precision in the field of control. Based on the characteristics of pattern recognition, control theory and intelligent manufacturing, we propose a more efficient control system design with Kinect, apply it to the vehicle control project, and get better results. In this paper, the Kinect equipment is applied to design a remote multi-person control program. By recognizing the skeleton information of Kinect in the software Processing, we combine Kinect with the skeleton extraction algorithm Alphapose to optimize the recognition process, and finally controls the vehicle designed with Arduino UNO board. Meanwhile, we make the prospect of applying the scheme to the actual project and put forward a new idea for the design of control systems.