The present invention relates to a method and a device for voice synthesis of an electronic book or an electronic document having a complex multi-layered structure. The present invention can be realized by connecting a desktop publishing (DTP) unit (102), a manufacturing unit (103), and a reading unit (104) to each other through a wired and wireless network (100) including an internet communication network. The DTP unit (102) completes a design by sequentially arranging layers of a text frame to which a text is inputted in a reading order from a lowest layer to a highest layer and converts the completed design into PDF and storing the PDF, and then uploading the PDF to a web server. The manufacturing unit (103) downloads the PDF to sequentially extract and merge text data and coordinate values from the lowest layer to the highest layer and stores them as a file in a XML format, thereby uploading the file to the web server or storing the file in a database. The reading unit (104) opens the file in the XML format uploaded to the web server or accesses the database to open the database file and extracts the text data to be read and stored in a variable, and, when read in a unit of a sentence divided by a period, performs voice reading by calling a TTS function of a TTS unit using the variable as a parameter or voice reading to a voice output unit by outputting as a message box or HTML format on a display unit, thereby synthesizing the voice in the TTS unit. According to the present invention, although the electronic book and the electronic document having the complex multi-layered structure for people without disabilities is not re-processed for a visually handicapped person only with additional time, efforts, and costs, the sentences can be sequentially voice-read from the TTS unit so that the visually handicapped person can easily understand a context and meaning thereof.COPYRIGHT KIPO 2017본 발명은 복잡한 다단 구조의 레이아웃으로 구성된 전자책 또는 전자문서의 음성 합성 방법 및 장치에 관한 것이다.본 발명은 텍스트가 입력된
