This　paper　constructs　a　stochastic　fuzzy　controller　based　on　OCPFA　learning　system　to　realize　self-balancing　control　of　two-wheeled　robot.　The　OCPFA　learning　system　is　in　fact　a　Probabilistic　Finite　Automata　(PFA)　which　based　on　Skinner　Operant　Conditioning　(Skinner　OC).　Reorientation　mechanism　which　take　for　posture　balance　of　orientation　function　as　goal-orientation　is　designed　to　make　response　to　the　output　control　variable　of　fuzzy　stochastic　controller;　Learning　mechanism　is　designed　to　update　probability　of　output　control　variable　by　using　the　response　information　from　the　environment　to　achieve　the　anticipant　probability　vector　which　can　minimize　orientation　function.　The　designed　stochastic　fuzzy　controller　can　choose　the　optimal　control　variable　by　interacting　with　the　dynamic　environment.　The　simulation　indicate　that　the　stochastic　fuzzy　controller　successfully　applied　in　two-wheeled　robot　self-balancing　without　requiring　the　model　of　the　robot　and　the　robot　can　show　control　behavior　of　autonomous　learning　which　similar　to　animal＇s　OC　learning　behavior.