In　wireless　sensor　networks　(WSNs),　mobile　agent　(MA)　can　effectively　gather　and　aggregate　the　sensed　data　so　as　to　minimize　the　network　energy　consumption.　In　some　realistic　applications,　the　aggregation　cost　cannot　be　neglected　and　MA　may　get　heavier　from　the　data　it　accumulated　as　it　migrates　through　the　touring,　which　easily　leads　to　the　high　communication　and　aggregation　cost　in　some　nodes.　In　this　paper,　we　first　propose　a　data-gathering　algorithm　which　is　optimized　by　discrete　particle　swarm　optimization　(DPSO)　based　on　MA.　In　the　itinerary　of　the　MA,　we　introduce　adaptive　data　aggregation　scheme　and　data　separation　scheme　to　reduce　the　energy　expenditure.　In　addition,　we　further　study　how　to　adaptively　adjust　route　structure　by　considering　the　remaining　energy　of　the　nodes　to　avoid　the　nodes＇　premature　death　for　extending　the　network　lifetime.　Simulation　results　show　that　our　algorithm　has　effectively　lower　energy　consumption,　higher　robust　structure,　and　longer　lifetime.