A generalized Bouc-Wen (GBW) hysteresis model is proposed to accurately characterize the hysteresis nonlinearity of the piezoelectric actuator, since the classical Bouc-Wen hysteresis model cannot accurately characterize the inherent asymmetric frequencydependent dynamic hysteresis nonlinearity of piezoelectric actuators. Firstly, based on the classical Bouc-Wen hysteresis model, two asymmetric terms and a second-order IIR filter are introduced to characterize the asymmetric hysteresis and high-frequency phase hysteresis of the piezoelectric actuator, and further analyze the model parameter values with respect to the frequency variation law to determine the frequency-dependent parameters of the model. Then, the experimental platform of precision positioning of the piezoelectric actuator based on NI CompactRIO measurement and control system is built, and the parameters of the GBW model are identified by the particle swarm optimization algorithm and the proposed GBW model is evaluated experimentally. Experimental results show that the maximum error of the GBW model is 0. 190 6 μm and the root mean square error is 0. 043 1 μm for the variable frequency sinusoidal excitation signal, which is only 0. 65% of the displacement range of the piezoelectric actuator, with a decrease of 82. 07% and 62. 10% compared to the classical Bouc-Wen (CBW) model and the enhanced Bouc-Wen (EBW) model, respectively. Compared with the CBW model and the EBW model, the proposed GBW model significantly improves the model accuracy and broadband performance, and the existence of the analytical inverse model is easy for controller design, which helps to realize broadband and high-speed precision positioning of the piezoelectric actuator in ultra-precision instruments and equipments.