The disclosure relates to a multi-audio-track sequence generation method, device and equipment and a storage medium, and belongs to the technical field of the internet. The method comprises the stepsof: generating a first melody unit sequence corresponding to an original audio; acquiring a preset database, wherein the preset database comprises a musical knowledge rule and a plurality of pieces oftemplate data; according to a matching degree between the first melody unit sequence and template data, acquiring template data matched with the first melody unit sequence from the preset database; according to the template data matched with the first melody unit sequence, obtaining a second melody unit sequence; and carrying out synthesis on the first melody unit sequence and the second melody unit sequence to obtain a multi-audio-track third melody unit sequence. According to the disclosure, the multi-audio-track melody unit sequence matched with the original audio is generated by utilizingthe musical knowledge rule so as to obtain a musical work matched with the original audio, improve the diversity and the expressive ability of the musical work, enhance interestingness and meet demands of users better.