The problem of discord detection in a time series has recently attracted much attention and several algorithms have been developed to tackle this problem. However, most of them suffer from high computational cost and hence can not suit real world applications well. In this paper, we propose a novel discord discovery algorithm, named Hash_DD, which is based on SAX representation and hashing. In comparison with HOT SAX, one of the most popular time series discord discovery algorithms, our hash-based algorithm accelerates the discord discovery process remarkably as well as reduces the memory cost. The experimental results have demonstrated that the proposed approach can not only effectively find discords in time series, but also greatly improve the computational efficiency.