This paper introduces the concept of input-to-output practical stability (IOpS), a generalization of input-to-state stability (ISS). It establishes two key results: first, that the interconnection of two IOpS systems remains IOpS if the composition of their gain functions is less than the identity function; second, that feedback can be used to assign gain functions. The paper applies these results to global asymptotic stabilization via partial-state feedback for linear systems with nonlinear, stable perturbations and systems with a particular disturbed recurrent structure. The paper discusses uncertain dynamical systems and presents a new design tool for stabilizing systems with complex structures. It proves that a certain uncertain dynamical system can be robustly stabilized using partial-state feedback. The system is described by a set of differential equations involving measurable state components and unmeasured components, with the unmeasured components satisfying input-to-state stability (ISS) conditions. The paper uses the ISS property introduced by Sontag to address dynamic uncertainties and systems with complex structures. It generalizes the "adding one integrator technique" and shows how the ISS property can be propagated through integrators to design stabilizing partial-state feedback. It also relates the IOpS concept to the classical input-output L∞ operator approach and to the "topological separation" concept. Section 2 introduces IOpS and establishes a generalized small-gain theorem, which extends a result by Mareels and Hill on monotone stability. It also discusses gain assignment by feedback. Section 4.3 shows how the ISS property can be used to design stabilizing partial-state feedback for the system. Section 5 contains the proofs of the main theorems. The paper uses standard notations, including the Euclidean norm, the identity function, and definitions of measurable input functions and their truncations.This paper introduces the concept of input-to-output practical stability (IOpS), a generalization of input-to-state stability (ISS). It establishes two key results: first, that the interconnection of two IOpS systems remains IOpS if the composition of their gain functions is less than the identity function; second, that feedback can be used to assign gain functions. The paper applies these results to global asymptotic stabilization via partial-state feedback for linear systems with nonlinear, stable perturbations and systems with a particular disturbed recurrent structure. The paper discusses uncertain dynamical systems and presents a new design tool for stabilizing systems with complex structures. It proves that a certain uncertain dynamical system can be robustly stabilized using partial-state feedback. The system is described by a set of differential equations involving measurable state components and unmeasured components, with the unmeasured components satisfying input-to-state stability (ISS) conditions. The paper uses the ISS property introduced by Sontag to address dynamic uncertainties and systems with complex structures. It generalizes the "adding one integrator technique" and shows how the ISS property can be propagated through integrators to design stabilizing partial-state feedback. It also relates the IOpS concept to the classical input-output L∞ operator approach and to the "topological separation" concept. Section 2 introduces IOpS and establishes a generalized small-gain theorem, which extends a result by Mareels and Hill on monotone stability. It also discusses gain assignment by feedback. Section 4.3 shows how the ISS property can be used to design stabilizing partial-state feedback for the system. Section 5 contains the proofs of the main theorems. The paper uses standard notations, including the Euclidean norm, the identity function, and definitions of measurable input functions and their truncations.