Accurate　estimation　of　State　of　Charge　(SOC)　based　on　equivalent　circuit　model　(ECM)　of　Lithium-ion　batteries　(LIBs)　is　an　important　research　topic.　Recent　research　has　found　that　fractional-order　model　(FOM),　as　one　kind　of　equivalent　circuit　model,　can　provide　a　better　description　of　LIBs　dynamics　than　the　conventional　integerorder　ECMs.　However,　it　is　difficult　to　directly　identify　an　FOM　online.　In　this　paper,　a　synergy　of　Beetle　Antennae　Search　and　Recursive　Least　Squares　(BAS-RLS)　is　proposed　to　identify　the　parameters　of　an　FOM　for　LIBs.　Specifically,　the　BAS　is　adopted　to　determine　the　fractional　order,　while　the　remaining　parameters　of　FOM　are　estimated　via　the　well-established　RLS　algorithm.　A　comparison　study　will　demonstrate　that　this　method　offers　a　similar　modeling　accuracy　as　the　Particle　Swarm　Optimization　(PSO),　and　the　online　estimated　model　is　more　accurate　than　offline　estimated　models.　Based　on　the　identified　FOM,　this　article　then　explores　the　battery　SOC　estimation　from　two　aspects:　initial　value　effect　of　the　SOC　and　the　comparison　of　multiple　estimation　algorithms,　thus　verifying　the　advantage　of　the　proposed　method　for　LIBs　modeling　and　SOC　estimation.