<p>This study investigates the fully distributed formation-containment control problem for a class of nonlinear multiagent systems (MASs) under directed communication topology. To solve this problem, a fully distributed adaptive formation-containment control approach incorporating neural network-reliable state constraints is developed. Specifically, a series of fully distributed observers is constructed for each agent to estimate the exosystem states, thereby eliminating the requirement for restrictions on the exosystem matrix and prior knowledge of the exosystem dynamics. Then, by incorporating neural network techniques and nonsmooth feedback, a novel adaptive formation-containment control protocol is designed for nonlinear MASs to achieve asymptotic tracking of the corresponding observer’s output. Furthermore, the proposed control protocol actively guarantees that agents’ states remain confined within user-defined compact sets, thereby enhancing the reliability of the neural network approximation. Under the proposed control approach, the leaders asymptotically achieve a time-varying formation, while the followers ultimately enter the convex hull constructed by the leaders. Finally, numerical simulations and analogue experiments applied to uncrewed ground vehicles are conducted to validate the effectiveness of the proposed approach.</p>