Embodiments of the present disclosure disclose a method and apparatus for recommending a video. The method includes: receiving a video recommendation request sent by a terminal device, the video recommendation request including a first voice, the first voice being a voice inputted by a user requesting a video recommendation; determining, from user voices stored in a server, a second voice having a greatest similarity with the first voice; and sending information of a target video to the terminal device according to a user profile corresponding to the second voice, if a first confidence recognizing the user as a user corresponding to the second voice being greater than or equal to a first threshold. The method and apparatus for recommending a video according to the embodiments of the present disclosure have a high accuracy of video recommendation.