Buckling　-restrained　braces　(BRBs)　can　mitigate　the　uneven　inter　-story　drifts　of　vertically　irregular　steel　frame　structures　during　earthquakes,　preventing　unfavorable　seismic　forces　and　failure　patterns.　However,　the　structural　lateral　displacement　profile　of　vertically　irregular　steel　buckling　-restrained　braced　frame　structures　(VSBFSs)　can　hardly　be　predicted;　hence,　parameter　tuning　and　iterative　designing　are　commonly　required　in　conventional　direct　displacement　-based　design　(DDBD)　methods,　which　strongly　rely　on　the　experience　of　designers　and　consumes　a　considerable　amount　of　time.　This　study　proposes　a　substructure　-based　design　(SBD)　method　developed　explicitly　for　VSBFSs.　In　this　method,　the　substructures　of　every　story　of　a　VSBFS　are　separated　from　the　entire　structure　and　individually　analyzed　in　the　subsequent　procedures.　The　boundary　conditions　of　the　substructures　are　initially　simplified　as　hinges,　and　an　explicit　equation　of　BRB　stiffness　and　inter　-story　drift　for　the　substructures　is　established.　An　iterative　calculation　procedure　is　consequently　applied　to　correct　the　equation　by　considering　the　influence　of　the　constraints　of　adjacent　stories.　In　this　manner,　the　inter　-story　drifts　of　all　the　stories　of　the　entire　structure　uniformly　converge　to　the　predetermined　targets,　in　which　the　demand　of　BRBs　for　each　story　is　also　determined.　Compared　with　conventional　DDBD　methods,　the　proposed　method　does　not　assume　the　structural　lateral　displacement　profile　and　is　not　based　on　experience,　yielding　relatively　economical　and　safe　design　outcomes　in　a　timely　manner.　The　proposed　SBD　method　was　first　validated　by　comparing　it　with　existing　elastic　and　elastoplastic　design　methods　through　a　15　-story　VSBFS　example.　Then,　the　applicability　of　the　SBD　method　was　further　explored　through　four　6　-story　VSBFSs　with　different　plan　layout　reductions　and　two　VSBFSs　with　heights　of　20　stories　and　30　stories.　Finally,　the　SBD　method　was　applied　to　redesign　the　BRBs　of　a　vertically　irregular　thermal　power　plant　model　(an　example　of　a　VSBFS)　in　a　shaking　table　test.