Communicating in short messages, such as using microblogs, was becoming more popular currently. Twitter https://twitter.com supports microblogs and retrieval of the blogs by users. To retrieve Twitter documents, we need specific strategies due to its specific characteristics. One new strategy for improving the effectiveness of twitter document retrieval is using the query expansion technique. This paper elaborates query expansion in twitter document retrieval by using the hashtag. We compared the effectiveness of query expansion in four different scenarios: the baseline result using no query expansion, highest scoredterm in terms of frequency-inverse document frequency (tfidf), maximum hashtag occurance, and combination of the highest scored-term and the maximum hashtag. The results show that the combination of the maximum term in tfidf and the maximum hashtag performs better in retrieving relevant documents than the baseline.