This paper presents a coordinated tracking control scheme of multi-agent systems (MAS) based on a data-driven approach. Each agent uses the golden section adaptive control as one of the data-driven methods to realize intelligent self-adaptive, and is able to quickly respond to the changing given target. Several proportional, integral and derivative (PID) controllers are applied as the coordinated controller along with improved adjacent coupling structure, simpler coordination structure and better synchronization performance are accomplished. The consensus and stability of closed-loop MAS using the proposed control scheme is analyzed by using a kind of necessary and sufficient conditions. Validity and feasibility of the coordinated tracking control scheme of MAS is verified by simulation examples.