This　article　was　based　on　the　application　of　music　innovation　and　second　creation　under　the　digital　age.　Through　the　statistical　analysis　of　the　user＇s　operation　record　of　the　music　and　predicting　the　music　audition　amount　of　the　artist　in　the　next　stage,　it　can　be　judged　that　the　music　audition　amount　was　the　highest　among　the　artists　in　a　period　of　time,　and　these　artists　represent　the　music　trend　in　a　period　of　time.　This　paper　predicted　the　amount　of　song　trial　of　the　artist　by　statistical　analysis　of　the　user＇s　audition,　download　and　collection　of　songs　recorded　by　the　electronic　music　platform　combined　with　the　second　exponential　smoothing　method,　the　autoregressive　moving　average　model　and　the　BP　neural　network　model,　At　the　same　time,　the　prediction　system　of　music　pop　trend　based　on　BP　neural　network　algorithm　was　designed　and　implemented.　©　2018　Walter　de　Gruyter　GmbH.　All　rights　reserved.