Tropicalcycloneorstormnearthecoastgeneratesalocalriseinsealevel,calledstormsurge.Therearetwogeneralapproachesthathavebeenusedtoestimatesealevelinstormconditions
statisticalmodelingandnumericalmodeling.TheaimofthispaperistofindaneffectivemodelbetweenMIKE21numericalmodelandanArtificialNeuralNetwork(ANN)modelinpredictingsealevelheightinQuiNhon,VietnamduringfourstormeventsKetsana-2009,SonTinh-2012,Nari-2013andWutip2013.Theresultsfromthetwomodelswerecomparedwithobservedsealeveldatainordertoevaluatetheirrespectiveperformances.TheresultsfurtherindicatethatthepredictionsfromANNandMIKE21modelmatchwellwiththeobserveddata.Theverificationshowsthattheneuralnetworkmodelshavethepotentialforsuccessfulapplicationinlocalwaterlevelforecastingsystems