A class of generalized Cohen-Grossberg neural networks(CGNNs) with variable delays are investigated. By introducing a new type of Lyapunov functional and applying the homeomorphism theory and inequality technique, some new conditions are derived ensuring the existence and uniqueness of the equilibrium point and its global exponential stability for CGNNs. These results obtained are independent of delays, develop the existent outcome in the earlier literature and are very easily checked in practice. §1 IntroductionCohen and Grossberg [1] proposed a class of neural networks in 1983, which can be described by the following ordinary differential equations:where n denotes the number of neurons in the network, x i (t) denotes the state of the ith neuron and the activation function f j (·) denotes the output of the jth neuron at time t, the n × n connection matrix W = (w ij ) represents the neuron interconnections. Model (1) includes a large number of models from neurobiology and population biology [2] . In particular, it includes the popular Hopfield neural networks [3] and BSB models. Moreover, this type of neural networks has been widely studied due to its potential applications in many areas such as associative memory and optimization. Because of the finite switching speeds of neurons and amplifiers, time delays inevitably exist in biological and artificial neural networks. Therefore, Cohen-Grossberg neural networks with time delays have been widely studied, which is described asReceived: 2008-03-23 MR Subject Classification: 34K20, 92B20, 37N25, 93C23