Real-time　hybrid　testing　(RTHT)　has　been　developed　rapidly　for　seismic　testing.　When　energy　dissipation　structure　is　tested　by　RTHT,　the　damper　is　usually　adopted　as　physical　substructure　and　the　structure　is　adopted　as　numerical　substructure.　In　order　to　minimize　the　seismic　response　of　energy　dissipation　structure,　many　dampers　installed　at　different　positions　are　often　necessary.　The　dynamic　response　of　dampers　at　different　positions　is　different　in　seismic　action.　Therefore,　more　than　one　damper　needs　to　be　tested　physically　in　RTHT,　and　multiple　dampers　loaded　by　multiple　actuators　in　real　time　will　increase　control　difficulty.　To　solve　this　issue,　an　offline　hybrid　testing　method　for　energy　dissipation　structure　is　proposed　based　on　machine　learning,　which　avoids　the　challenge　of　control　and　stability　caused　by　real-time　loading　of　multiple　actuators　in　RTHT.　The　feasibility　of　this　method　is　proved　by　numerical　simulation,　which　provides　a　new　idea　for　seismic　testing　of　energy　dissipation　structure.　©　2023,　The　Author(s),　under　exclusive　license　to　Springer　Nature　Singapore　Pte　Ltd.