Based　on　reinforcement　and　dynamic　programming,　a　novel　adaptive　control　algorithm　is　proposed　for　the　complex　systems　which　have　the　characteristics　of　model　unknown,　speediness,　multiple　variables,　serious　nonlinear.　The　method　needs　not　know　the　priori　knowledge　of　system,　uses　a　double　BP-network　to　control　system　with　hierarchical　theory.　combines　the　self-tune　property　of　reinforcement　learning,　and　at　last　effectively　controls　the　unstably　nonlinear　system.　The　paper　experimental　object　is　a　single　inverted　pendulum.　It　is　shown　from　the　simulation　results　that　this　method　has　good　control　effect,　good　steady　accuracy　and　good　interference　rejection.