Disclosed are a method and device for processing video data. The method comprises: an encoding end encoding acquired original video data according to Hierarchical P-frame Prediction (HPP) to obtain an HPP code stream; conducting redundant encoding on the HPP code stream according to forward error correction (FEC) encoding, wherein the number of redundant packets of each frame in the HPP code stream progressively decreases according to the time layer where each frame is located in the HPP in the order of lowest to highest; and finally sending each frame in the HPP code stream after the redundant encoding to a decoding end in sequence. The present application combines an HPP encoding structure and FEC redundant encoding, and adds a redundant packet to a HPP code stream encoded according to the HPP encoding structure. In this way, even though some packet loss occurs during data transmission, the decoding end can still perform lost packet recovery through the redundant packet, greatly improving the success rate of decoding, and reducing the stutter time of video.