In this paper, Hopfield network is introduced for the restoration of extracted watermark image which may be blurred due to the signal transfer or various signal processing operations. A novel codebook method is designed to reduce the storage space of the network and to increase the security. First, each watermark image is divided into adjacent and non-overlapped sub-block images and mapped into a codebook. Second, this codebook is encrypted by a chaotic sequence. During the process of watermark restoration, the codebook can be obtained via a secret key which is then used to construct block weight matrix of the neural network for the restoration of the blurred watermark images. Simulation results demonstrate the excellent performance of the proposed method.