The invention discloses an audio and video subtitle generation method and apparatus. The method comprises the following steps: receiving voice data with subtitles to be generated and text data; according to the rhythm of speakers, segmenting the vice data to obtain each voice data segment according with habits of the speakers; according to the voice data segments, segmenting the text data to obtain text data segments corresponding to the voice data segments; and according to time information of each voice data segment, obtaining starting time and finishing time of the text data segments corresponding to the voice data segments. According to the invention, the text data can be displayed synchronously with the voice data in a simple and efficient mode, and the generated subtitles are more complete.