Aiming　at　the　time　lag　of　structural　vibration　control　and　the　demand　of　feedback　control　in　full　structural　state,　a　sequential　optimal　predictive　modal　control　(SOPMC)　algorithm　based　on　state　observer　was　proposed.　According　to　the　structural　response　of　feedback,　the　full　state　of　structure　was　observed　with　the　state　observer,　and　the　real-time　control　force　was　calculated　with　the　SOPMC　algorithm.　The　radial　basis　function　neural　network　(RBFNN)　was　designed　through　taking　the　structural　response　of　feedback　and　the　calculated　control　force　as　input　and　with　considering　the　time　lag　of　control　system,　and　then　a　sequential　optimal　predictive　modal　controller　based　on　full-state　observer　was　developed.　A　seismic　response　control　was　carried　out　for　an　actual　vibration　isolated　330　kV　voltage　instrument　transformer　with　distribution　parameters.　The　simulation　results　show　that　without　the　presence　of　time　lag,　the　control　effect　of　SOPMC　algorithm　is　equivalent　to　that　of　sequential　optimal　modal　control　(SOMC)　algorithm.　With　the　presence　of　time　lag,　the　SOPMC　algorithm　can　effectively　overcome　the　influence　of　time　lag　on　the　optimal　predictive　control　algorithm,　and　exhibits　good　control　effect.