Development of smart materials with inherent damage sensing capabilities is of great interest to aerospace and other structural applications. Most of the existing smart materials are based on using embedded sensors for structural health monitoring. However, embedded sensors can lead to undesirable effects such as stress concentration and can cause premature failure. Therefore, using microstructural components for additional function of sensing of the structural health is the only option. Such possibilities exist only in selected few materials. The present study investigates the feasibility of developing fiber- and particle-reinforced composites into smart materials. The sensing approach considered is based on the morphology-dependent shifts of optical modes, referred to as the whispering gallery modes (WGMs), of spherical dielectric micro-particles. The WGMs are excited by coupling light from a tunable diode laser using single mode fibers. The WGMs of the micro-particles can be observed as sharp dips in the transmission spectrum through the fiber and are highly sensitive to the morphology of the particle. A minute change in the size, shape, or refractive index causes a shift of the optical modes, which can be interpreted quantitatively in terms of the parameter that caused the change. A theoretical framework is developed for such sensor systems that provides quantitative relations between the stress applied on the micro-particles and corresponding shift in WGMs. These relations are validated against the available experimental results.