This paper investigates the formation control problem of multiple mobile agents with second-order nonlinear dynamics in complex environments containing multiple obstacles. By employing the null-space-based behavioral (NSB) control architecture, a novel fast terminal sliding mode based adaptive control strategy is designed combining the fuzzy logic systems (FLS), which can guarantee each individual agent to achieve the predesigned desired position within the formation, and more importantly guide the overall systems to achieve desired formation while avoiding obstacles. Finally, simulation examples demonstrate the effectiveness of the algorithms, performed by Matlab numerically and Pioneer 3-AT mobile robots in practice, respectively.