This　article　focuses　on　the　most　common　application　scenarios　for　data　collection　and　uploading　inWSN(Wireless　Sensor　Networks).　First,　we　measure　the　energy　consumption　of　widely　used　hardware.　According　to　the　characteristics　of　transmission　energy　consumption,　a　MIP　(mixed　integer　programming)　model　called　FAT-WSN　(fragmentation　aggregation　transmission　WSN)　is　proposed　to　minimize　the　number　of　data　fragments.　Moreover,　we　propose　an　iterative　solution　for　this　MIP　problem　with　elasticity　and　low　complexity.　The　main　optimization　method　for　this　model　is　to　adjust　topology　and　traffic　distribution.　It　focuses　on　optimizing　the　number　of　data　transfers　without　modifying　any　data　and　without　introducing　a　compression　calculation　burden.　Finally,　simulation　and　small-scale　real　node　verifications　are　performed　for　the　FAT-WSN　scheme.　The　experimental　results　show　that　FAT-WSN　can　effectively　reduce　the　number　of　data　transmission　and　reception,　thereby　reducing　energy　consumption　and　improving　network　life.　Compared　with　the　MinST　model,　JGDC　(Jointly　Gaussian　Distributed　Compress)　model　and　AMREST　(Approximately　Maximum　min-Residual　Energy　Steiner　Tree)　model,　the　network　life　can　be　increased　by　10%-30%　without　extending　the　calculation　time.