This paper studies observer-based neural control of hypersonic morphing vehicles using experience replay enhancement. The longitudinal model of the aircraft is divided into a velocity subsystem and a height subsystem. For the velocity subsystem, an observer-based dynamic inversion control scheme is designed considering complex external disturbances. For the altitude subsystem, a neural backstepping control scheme is designed by combining the experience-replay-based observer. The major merit lies in the neural observer design using experience replay enhancement, distinct from conventional disturbance observer that only depends on the current state information obtained, may failing in addressing the coupling effects caused by morphing-induced uncertainties and unsteady aerodynamics. Numerical simulations validate the effectiveness of the proposed control.