Abstract: The global asymptotic synchronization of a class of delayed chaotic neural networks is studied on the basis of drive-response synchronization. Constructing a suitable Lyapunov-Krasovskii functional and based on the scheme of decentralized control, the design of a control law is proposed to ensure the global asymptotic synchronization of state trajectories of two chaotic neural networks of which the structure are the same and the initial conditions are different. In this way the control law obtained is not only easy to be implemented but highly reliable in practice, thus making up for the inadequacy of conventional centralized control. Two illustrative examples are used to demonstrate the effectiveness of the proposed method.