We propose a mathematical model and a recurrent algorithm to perform the automatic navigational markup of audiobooks by their exact text electronic equivalent. This approach allows to convert unmarked audiobooks into DAISYbooks for people with visual impairments. A speech synthesizer converts each markup fragment of the text of the book into an audio signal. We search an audio fragment closest to the speech synthesizer audio signal in specific borders of the audiobook using the Dynamic Time Warping (DTW) algorithm. As a result, we automatically obtain the markup of the audiobook. We investigated statistical characteristics of markup errors.