This paper proposes the intelligent system to assist the visually-impaired people in reading printed Chinese document. As the document may consist of the combination of text and photos, the proposed system first applies the text frame detection scheme to identify the text parts of the document, and utilizes the Optical Character Recognition (OCR) to convert the text image into pure text file. The aim of this paper is to implement the automatic text summarization algorithm to get the brief meaning of the whole text. And the visually-impaired people can listen to the summary of the document through the existing text to voice tool. In the system, the neural network model is adopted to find out text blocks before recognizing text, and to optimize the Chinese automatic text summarization system. The performance of the proposed scheme is also evaluated by the TextRank with Word2Vec. The proposed system can assist the visually-impaired people especially in daily life when it is applied in the wearable devices.