Active vibration control is not only suitable but also urgently needed in order to mitigate excessive vibration for dynamic systems such as wind turbine towers. However, there is a challenge of directly observing displacement and developing an efficient reduced-order control strategy. With this in view, a reduced-order controller based on displacement observation is developed for the vibration mitigation of flexible high towers, where the problems of low signal-to-noise ratio (SNR) and intricate acceleration-based control algorithms are eliminated completely. The displacement can be measured by laser or modern videometrics technology. A reduced-order model with an active mass damper (AMD) at the top of the tower is created using modal truncation approach based on modal participation factor (MPF), which successfully preserves the dominant modes of the original structure, and reduces the influence of high-frequency uncertainties. Considering structure uncertainties caused by environmental factors and aging process, an unscented Kalman filter (UKF) is introduced to identify the modal stiffness of the reduced-order model. The present study, using a 5 MW onshore wind turbine as an example, investigates active vibration control under wind and seismic loads and compares it with traditional tuned mass damper (TMD). From a dynamic perspective, the simplified model of wind turbine tower takes the form of a multi-degree-of-freedom series model. Consequently, a six-story steel frame platform has been constructed to experimentally validate the efficacy of the reduced-order controller. According to numerical simulation and laboratory testing, the control effect of the reduced-order controller with a few displacement states as feedback can be utilized as an active control strategy for high-rise flexible towers.