We use feature-based modeling framework for control of loads on aeroelastic wings in the presence of unsteady wind gust disturbances, leveraging interpretable reduced-order-model system identification techniques without relying on black-box machine learning techniques. Unlike prior formulations, we explicitly include the dominant disturbance, such as a gust, as a control input within the model, allowing the controller to respond adaptively and in anticipation to external forcing. To this end, model predictive control (MPC) could be used in the low dimensional latent space of features, whose dynamics are identified partly by sparse identification of nonlinear dynamics with control (SINDYc) and linear parameter-varying system (LPV). As a proof of concept, the methodology is applied to system identification of smart vortex generators (SVGs) for mitigating transient gust loads on an aeroelastic wing section in CFD simulations. This methodology offers a promising path toward real-time mitigation of atmospheric disturbances in next-generation flight systems.