This article analyzes exponential synchronization for a class of Cohen-Grossberg neural networks with time-varying delays. Firstly, according to the concept of synchronization, a controlled response system is constructed, and the error system is obtained. Secondly, by establishing suitable Lyapunov functions and using inequality techniques, sufficient conditions for exponential synchronization of the error system under different controllers are provided, and the exponential convergence rate of the system is given. Finally, two examples are used to verify the effectiveness of the theoretical results.INDEX TERMS Cohen-grossberg neural network, time-varying delays, exponential synchronization, impulsive control.