System modeling with respect to reliability function is of increasing importance, since knowledge about reliability function enables its integration in control affecting the control strategies. Suitable operating strategies will lead not only to extended lifetime, but also give a possibility for predictive maintenance. The assessment of reliability function relies often on probabilistic theory. From probabilistic point of view it consequently involves mathematical assumptions, but physically mainly relies on knowledge about systemÕs state-of-deterioration. This implies that statements about the state-of-deterioration, about predictive maintenance, as well as the ability to control the systemÕs extended lifetime assume knowledge about the relation between systemÕs load (applied to the system), systemÕs stress (appearing within the system), and the resulting behavior of systems to the loads applied. This contribution establishes a parametric model from the experimental data generated from wear experiments, and compares the results to those results obtained for individual system generating data.