This invention discloses a method and apparatus for
processing continuous audio streams in storage-constrained recording devices, belonging to the field of audio
data processing technology. The method is executed by the recording device and includes: encoding and segmenting the received continuous audio
stream to obtain multiple audio data blocks, and determining the current audio block to be stored from the obtained audio data blocks; acquiring a first available storage space and a used storage space in the local audio buffer; if the first available storage space is detected to be lower than a first threshold, or the used storage space is detected to be higher than a second threshold, filtering out the audio block to be deleted from the segmentation index table according to the audio data block information recorded in the segmentation index table; deleting the audio block to be deleted from the local audio buffer, and acquiring a second available storage space in the local audio buffer at this time; and storing the current audio block to be stored according to the second available storage space. This invention improves the reliability and availability of the recording device.