This paper describes the design of an output-feedback controller based on an adaptive fuzzy observer for uncertain single-input, single-output nonlinear dynamical systems. We have focused on the realization of a minimal dynamic order for the adaptive fuzzy observer. For this purpose, we adopt an adaptive fuzzy observer in which no strictly positive real (SPR) condition is needed and we combine a self-structuring scheme with an on-line estimation of fuzzy parameters. By using this proposed scheme, we can reduce the dynamic order of the adaptive output-feedback fuzzy control system. The Lyapunov synthesis approach is used to guarantee a global uniform ultimate boundedness property of the state estimation error and tracking error, as well as all other signals in the closed-loop system. No a priori knowledge of an upper limits on the uncertainties including optimal parameters and modeling errors is required. The theoretical results are illustrated through simulation examples.