Limited　progress　in　understanding　blast　mechanisms　has　led　to　significant　discrepancies　between　the　outcomes　of　existing　blasting　simulation　techniques　and　actual　blasting　results,　making　it　difficult　to　predict　muckpile　characteristics,　optimize　blasting　designs,　and　guide　on-site　production.　To　address　this　challenge,　this　study　presents　a　machine-learning-aided　(ML-aided)　method　for　blasted　muckpile　analysis,　based　on　an　innovative　ML-aided　post-blast　ore　boundary　determination　technique　developed　by　the　authors.　This　method　enables　accurate　calculation　of　muckpile　shape　and　ore　distribution　through　six　key　steps:　blast-induced　rock　movement　database　collection,　machine-learning　rock　movement　prediction　model　development,　prediction　model　performance　evaluation,　blast　block　meshing,　rock　movement　prediction,　and　rock　element　redistribution.　Using　this　approach,　the　blasted　muckpile　and　ore　recovery　can　be　predicted　in　advance.　In　this　study,　comparative　analysis　of　simulated　and　field　results　from　the　dividing　open-pit　blast　(DOPB)　and　center-initiation　open-pit　blast　(CIOPB)　methods　demonstrated　the　engineering　potential　and　effectiveness　of　this　approach　in　reducing　ore　and　profit　losses　through　improved　blasting　techniques.　This　method　can　be　regarded　as　a　valuable　addition　to　blast　simulation　techniques　and　highlights　the　potential　for　integrating　artificial　intelligence　in　mining　engineering.