This paper aims at proposing an adaptive neural networks control architecture for constrained bilateral teleoperation systems with delay. To achieve full-state constrained conditions, a cot-type barrier Lyapunov functions (BLFs) is built to analyze the stability of systems. The radial basis function neural networks (RBFNNs) are applied to estimate uncertainties in dynamic parameters and improve robustness of systems. Under the proposed control strategy, the stability of systems is proved, and the tracking performance is guaranteed considering full-state constraints and uncertainties. The simulation verifies the effectiveness of designed control scheme.