A voice output device, includes: a compound word voice data storage unit that stores voice data in association with each of compound words which is formed of a plurality of words; a text display unit that displays text containing a plurality of words; a word designation unit that designates any of the words in the text displayed by the text display unit as a designated word based on a user's operation; a compound word detection unit that detects a compound word in which voice data is stored in the compound word voice data storage unit, from among the plurality of words in the text containing the designated word; and a voice output unit that outputs voice data corresponding to the compound word detected by the compound word detection unit as a voice.