Trusted　and　reliable　wireless　sensor　networks　(WSNs)　rely　on　the　accurate　and　rapid　detection　of　anomalies.　However,　the　security　model　in　wired　networks　is　not　suited　for　WSNs　because　of　their　energy　constraints.　In　this　paper,　a　traffic　prediction　model　based　on　gene　expression　programming　(GEP-ADS)　is　proposed,　to　predict　the　time　series　of　normal　traffic.　Then　we　present　a　lightweight　anomaly　detection　scheme　(ADS)　in　WSNs.　In　the　ADS,　no　more　cooperation　between　sensor　nodes　is　required,　which　dramatically　decrease　the　energy　consumption.　Another　key　advantage　of　our　approach　is　that　GEP-ADS　could　solve　the　problem　that　the　traditional　time　series　methods　can＇t　make　an　accurate　prediction　without　the　pre-knowledge.　In　order　to　evaluate　our　ADS,　we　have　simulated　several　routing　attacks,　and　fully　discussed　the　parameter　setting　influence　on　our　ADS.　The　simulation　results　show　that　our　ADS　is　able　to　achieve　high　detection　accuracy　with　a　low　false　positive　rate.　©　2009　IEEE.