The invention discloses a flexible mechanical arm system fuzzy control method based on a series/parallel estimation model. The method comprises the steps of: building a dynamic model of a flexible mechanical arm servo system, and initializing a system state, a sampling time and control parameters; compensating unknown nonlinear uncertain items in the mechanical arm system by adopting the self-learning and self-adaptive capacity controlled by a fuzzy neural network; introducing virtual control quantities in each design through a dynamic surface technology, and preventing the complexity explosion problem caused by a traditional inversion control method through a first-order low-pass filter; and meanwhile, determining variable prediction values and designing related change rules based on the series/parallel estimation model to improve robustness of the control system. The fuzzy self-adaptive control method based on the series/parallel estimation model can effectively improve the control performance of the flexible mechanical arm servo system and realize quick and precise system tracking.