Left　ventricular　assist　devices　are　increasingly　used　for　long-term　support　in　heart　failure　patients.　It　is　important　to　find　an　optimum　operating　point　for　the　pump　that　is　appropriate　for　the　existing　function　of　the　heart　and　the　state　of　the　circulatory　system.　Therefore,　baroreflex　sensitivity　(BRS),　as　an　indicator　of　heart　function,　is　chosen　as　the　control　variable.　In　order　to　find　an　optimum　point　automatically,　an　extremum　search　algorithm　(ESA)　is　designed　to　find　an　optimal　mean　arterial　pressure　(MAP),　for　which　the　BRS　is　maximum.　Then,　a　MAP　controller　based　on　model-free　adaptive　control　is　designed　to　ensure　that　the　measured　MAP　tracks　the　desired　one.　In　order　to　test　the　feasibility　of　the　control　strategy,　numerical　simulations　and　simplified　in　vitro　experiments　were　conducted.　A　mathematic　model　of　the　cardiovascular　system　simulating　left　ventricular　failure,　physical　activity,　and　recovery　of　cardiac　function　is　used　in　the　simulation.　The　numerical　simulations　show　that　the　maximum　value　of　BRS　can　be　found　automatically　by　using　ESA.　The　rotational　speed　of　the　pump　is　automatically　increased　(from　6500　rpm　to　7000　rpm),　and　peripheral　resistance　is　decreased　to　simulate　slight　physical　activity.　When　E-max　is　increased　from　0.6　mm　Hg/mL　to　1.8　mm　Hg/mL　to　mimic　heart　recovery,　the　speed　is　decreased　from　7000　rpm　to　6300　rpm　in　response.　The　optimum　operating　point　for　the　pump　can　be　detected　by　the　proposed　control　strategy　without　the　need　to　set　a　reference　value　for　the　control　variable　by　operators.