Real-time　dynamic　substructure　(RTDS)　testing　combines　numerical　simulation　with　experimental　test.　The　constraint　between　the　two　parts　is　coupled　through　real-time　data　interaction,　so　that　the　dynamic　characteristics　of　the　tested　structures　can　be　uncovered　practically.　The　stability　of　RTDS　system　is　the　premise　of　a　successful　experimental　implementation,　which　determines　the　feasibility　of　RTDS　testing.　However,　the　stability　of　the　experiment　systems　may　be　affected　by　the　mass,　damping　and　stiffness　of　numerical　and　physical　substructure　and　the　characteristic　of　shaking　table　comprehensively.　Besides,　the　specimen-table　interaction　will　aggravate　the　coupling　effects　of　these　parameters.　Besides,　nonlinear　analysis　of　structures　is　of　the　utmost　significant　for　experiment　and　theoretical　researches.　The　comparison　between　linear　structure　and　nonlinear　structure　has　been　discussed　by　a　method　based　on　the　concept　of　＂gain　margin＂.　This　method　not　only　can　analyse　the　influence　of　a　single　parameter,　but　also　consider　the　coupling　effects　of　all　parameters.　The　influence　of　the　stability　for　shaking　table　based　RTDS　caused　by　specimen-table　interaction　was　analysed　numerically.　The　comparative　results　show　that　the　specimen-table　interaction　has　positive　influence　on　the　RTDS　stability　within　a　certain　range　and　has　negative　influence　in　other　ranges.　Therefore　the　influence　of　specimen-table　interaction　cannot　be　neglected,　and　it　needs　to　be　considered　additionally　for　a　successful　RTDS　implementation　in　advance.