The　clustering　analysis　of　cyclone　track　is　built　on　the　similarity　between　them.　Current　studies　on　trajectory　similarity　distance　were　for　one　or　two　attributes　only.　So　in　the　case,　the　results　were　unable　to　fully　reflect　the　motion　characteristics　of　trajectory.　In　order　to　make　full　use　of　the　trajectory　attributes,　based　on　the　grey　theory,　this　paper　proposed　that　the　three　attributes　(position,　direction　and　velocity)　should　be　fused　as　structure　distance　for　clustering　analysis,　helping　to　achieve　a　more　comprehensive　understanding　of　the　purpose　of　trajectory　motion.　Example　analysis　shows　that　the　algorithm　can　fully　reflect　the　trajectory　characteristics,　and　the　results　have　more　practical　significance.　©　2017,　Editorial　Board　of　Journal　of　Systems　Engineering　Society　of　China.　All　right　reserved.