This paper presents a translation dictionary expansion method based on transliteration. We have a Japanese-Uzbek dictionary, but it is not extensive enough for practical use. We also have a Japanese-Uighur dictionary. Since Uighur and Uzbek are similar languages, they have many similar words. So we can expand our Japanese-Uzbek dictionary using transliteration from Uighur to Uzbek. In this paper, we automatically collected 2,660 pairs of Uighur and Uzbek words similar to each other, and used them as training data for dictionary expansion based on transliteration. We tried two transliteration approaches; one uses handwritten rules and the other is a statistical one, and we show the feasibility of dictionary expansion based on transliteration.