With the revolution in technology and industry, the connectivity of electronic devices has gradually shifted from wired to wireless after long-time exploration. As a solution for power delivery, the non-contact power transfer holds promise charging for moving devices such as sensors, microcomputers, and robots, enabling battery-free sensing, processing, and communication. To reach the goal, we propose the adaptive wireless-powered network (AWPN) based on a full-duplex metasurface to realize a non-contact power supply for target tracking and wireless communications. The fabricated battery-free AWPN can obtain stable powers to perceive and compute the environmental data, which are then informed to the users by wireless communications. In particular, the proposed AWPN is good for moving devices, in which near-field positioning is achieved by the programmable metasurface combined with a convolutional neural network. AWPN can get more than 92% classification accuracy to provide precise positions of the moving targets for beam tracking. Thus, being adaptive and contactless, this AWPN will further propel the advancement of fields such as the Internet of Things (IoT), intelligent metasurface, and the robot industry.