In　this　paper,　a　novel　dimension　reduction　approach　with　the　application　of　grey　relational　analysis　based　on　difference　information　theory　is　proposed.　In　this　approach,　dimension　reduction　process　is　performed　based　on　the　calculation　of　the　similarity　between　each　pair　of　dimensions　with　the　application　of　grey　relational　analysis.　The　effectiveness　and　efficiency　of　the　proposed　approach　is　proved　by　the　experiments　on　real-world　datasets.　©2009　IEEE.