Designing automatic flight control systems for flexible aircraft is still an open and challenging task. The challenges range from theoretical-practical issues, such as dealing with uncertain, high-order, and increasingly complex models or overcoming the limitations and the difficulties imposed by most modern control techniques, to potentially catastrophic physical phenomena, e.g. aeroservoelastic instability, which is one of the major concerns in the design process. Traditionally, the design is based on rigid-body flight dynamics models with at most quasi-steady aeroelastic corrections, and aeroservoelastic stability problems are solved afterward by the addition of low-pass or notch filters in the feedback and feedforward loops. However, in more flexible airplanes this approach may be inadequate to filter elastic effects in the measured signals, resulting in low-performance controllers, poor handling and ride qualities, and even instabilities. Therefore, methodologies for flight control law design that take into account the aeroelastic dynamics are being pursued and correspond to the subject addressed in this thesis. Two output-feedback-based stability augmentation systems (SASs) are developed. The first one results from an original approach that combines the linear projective control technique with a coupled (aeroelastic + rigid-body) flight dynamics model of a flexible aircraft. It is shown that with suitable set of measurements such approach can easily incorporate frequency, damping, and dynamic decoupling requirements avoiding the conventional solutions based on filtering. Moreover, the resulting controller presents a simple structure (only a static output-feedback) and satisfactorily recovers robustness and performance characteristics derived from an optimal full state feedback solution. The second SAS corresponds to a Linear Quadratic Gaussian regulator augmented by an adaptive output-feedback control system to deal with model uncertainties. The latter is based on the well-established closed-loop reference model adaptive control technique for multi-input multi-output non-square systems. Its design stems from the Lyapunov direct method and the global stability is demonstrated by Lyapunov-like analysis using Barbalat's Lemma. The novelty here lies in the practical solutions proposed to properly handle the restrictions imposed by the technique. Both closed-loop stability and reference model performance are achieved even in the presence of uncertainties due to aircraft configuration changes, variations in the flight velocity and inaccurate effectiveness of the actuators. The drawback is the greater complexity and the higher order of the resulting controller. The X-HALE flexible experimental prototype, originally developed at the University of Michigan and that currently is also operated at the Instituto Tecnológico de Aeronáutica is considered in all case studies.