Recently, anonymization methods to protect data privacy have received a lot of attention. Most of them, however, focus on the anonymization methods for static tabular data. With the increasing use of the data, the privacy protection methods suitable for a variety of environments are required. In this paper, we propose an efficient anonymization method to protect the privacy in data streams. Since the previous work which considered the privacy protection in data streams is limited to the k-anonymity model, we develop the anonymization method for the ${ell}$-diversity model which guarantees stronger privacy protection. The proposed method considered the distribution of data streams and the frequency of sensitive information to minimize the information loss. Also, we performed various experiments to evaluate the efficiency and performance of the proposed method by using real data.