The　purposes　of　this　study　are　to　quantify　the　natural　resource　consumption　of　primary　aluminum　production　in　China　and　to　determine　the　resource-saving　potential　of　aluminum　in　the　transportation　sector　relative　to　steel.　In　this　study,　exergy,　which　expresses　both　the　quantity　and　the　quality　of　a　resource,　was　adopted　for　natural　resource　accounting.　The　cumulative　exergy　demand　(CExD)　of　primary　aluminum　was　calculated　by　the　process　analysis　method,　which　begins　in　the　final　link　of　the　fabrication　of　the　considered　product　and　runs　through　the　production　of　semifinished　products.　This　method　can　provide　detailed　information　about　each　process.　The　CExD　value　of　1　t　of　primary　aluminum　produced　in　China　is　246,778　MJ(ex),　and　the　largest　contributor　to　the　CExD　is　electricity,　which　is　mainly　consumed　in　the　electrolytic　process.　Compared　with　the　CExD　derived　from　the　national　average　resource　consumption　data　of　alumina　production　(three　techniques),　if　alumina　was　only　produced　by　the　Bayer　process,　the　CExD　of　primary　aluminum　would　decrease　by　10　%.　Taking　wheel　hubs　as　a　case　study,　although　aluminum　wheel　hubs　have　a　lower　natural　resource　consumption　rate　during　the　use　phase　relative　to　steel　wheel　hubs,　no　natural　resource　savings　are　obtained　before　a　certain　driving　distance　(breakeven　distance)　is　reached　because　of　higher　resource　consumption　in　the　production　phase.　The　total　amount　of　natural　resources　that　aluminum　wheel　hubs　could　save　(relative　to　steel　wheel　hubs)　over　its　lifetime　driving　distance　is　3,652　MJ(ex).　In　contrast,　the　breakeven　distance　derived　from　the　CML　model　is　approximately　80　%　lower　than　that　derived　from　the　CExD　model.　A　determination　of　the　advantages　of　the　Bayer　process　in　terms　of　resource　saving　suggests　that　importing　high-grade　bauxite　from　abroad　to　promote　the　application　of　the　Bayer　process　is　an　effective　way　to　reduce　primary　aluminum＇s　CExD,　i.e.,　resource　intensity,　in　China.　A　comparison　of　the　characterization　results　between　the　CML　model　(China＇s　own　factors)　and　the　CExD　model　shows　that　the　CExD　model　assigns　more　weight　to　coal　than　to　minerals,　whereas　the　CML　model　assigns　more　weight　to　minerals.