Localization　of　sensor　nodes　in　wireless　sensor　networks　(WSNs)　is　very　important　since　it　associates　spatial　context　with　the　data　collected　by　sensor　nodes　and　used　in　applications.　Rapid　development　of　wireless　sensor　technologies　and　wide　applications　of　wireless　sensor　networks　have　also　made　security　in　sensor　localization　a　primary　concern　as　well　as　a　great　challenge.　Without　adequate　security　measures,　the　performance　of　sensor　localization,　e.　g.,　the　accuracy　of　localization　results,　cannot　be　ensured　in　hostile　environments.　In　this　paper,　we　propose　a　trust-based　secure　sensor　localization　scheme　(TSLS)　for　WSNs　following　the　theory　of　neural　network　(NN).　The　proposed　TSLS　scheme　can　ensure　that　unknown　sensor　nodes　will　get　credible　information　to　perform　localization　through　the　evaluation　of　beacon　nodes　in　a　WSN.　The　evaluation　model　is　comprised　of　the　evaluation　of　both　the　identity　and　the　behavior　of　beacon　nodes　as　well　as　a　filtering　mechanism　to　deal　with　the　slander　behaviors　between　beacon　nodes.　Simulation　results　show　that　the　proposed　TSLS　scheme　can　improve　the　accuracy　of　sensor　localization　in　hostile　environments　for　both　static　and　dynamic　WSNs.