A continuous-time Hopfield neural network with two delay-connecting neurons is considered in this paper. Some sufficient conditions for the number and delay-independent stability of the equilibria in the network are given analytically. It is necessary to classify the attraction domains since multiple attractors coexist when the sufficient conditions are satisfied. Thus, effects of the delays on the boundary separating the basins of attraction of the stable equilibria are investigated analytically and numerically. The results show that the evolution of the boundary depends on the delays and is neither simple nor intuitive even if the delays do not affect the stability of attractors. The results provide also the possibility to design the network according to the memory pattern and storage.