An audio data parsing method, comprising: a parsing device receiving audio data (101) to obtain data pending parsing; revising the data pending parsing to obtain revised data pending parsing (105); conducting a first decoding for the revised data pending parsing to obtain a first decoded data (106); determining whether the decoded data pending parsing contains header data (107), and if yes, then obtaining data length, and obtaining a corresponding data unit (111) in the first data according to the obtained data length; when all corresponding data units are successfully obtained, conducting a second decoding for the obtained data units, obtaining the corresponding byte data according to the decoding result and checking the corresponding byte data, otherwise storing the received audio data in a first data buffer, and continuing to receive audio data. The method improves the audio data parsing efficiency and accuracy, and can be widely used in digital communication between both parties engaged in headset communication.