A　software-defined　wireless　sensor　network　can　dynamically　configure　the　nodes　in　a　network　according　to　the　demand　of　the　application　layer.　In　practical　applications,　such　as　environmental　monitoring,　the　nodes　in　a　wireless　sensor　network(WSN)　are　deployed　in　the　field　environment　on　a　large　scale,　and　the　data　rely　on　multihop　transmission　to　reach　the　sink　node.　The　data　extremely　easy　to　selective　forward-ing　attacks　during　data　transmission.　Therefore,　this　study　analyzes　the　models　of　selective　forwarding　attacks　and　proposes　an　abnormal　node　detection　method,　which　includes　a　node　behavior　measure-ment　scheme　and　trust-value　evaluation　mechanism.　In　addition,　the　application　of　a　software-defined　network　(SDN)　presents　increasing　network　delay.　Hence,　herein　a　network　recovery　mechanism　was　de-signed　based　on　cloud-edge　cooperation　to　ensure　the　rapid　recovery　of　the　network　after　identifying　the　abnormal　nodes.　Moreover,　experiments　were　conducted　using　simulation　software　and　actual　hardware.　We　verified　the　effectiveness　of　the　proposed　scheme.　The　experimental　results　revealed　that　the　pro-posed　method　can　effectively　identify　abnormal　nodes,　reduce　the　packet　dropping　ratio　and　shorten　the　network　recovery　delay　by　77.2%.　The　research　in　this　paper　solves　the　security　problem　of　SDWSN.(c)　2022　Elsevier　Ltd.　All　rights　reserved.