This paper focuses on observer design for the biologic system to estimate the sensitive population S. In fact, we synthesis two types of observer: the first one concerns this nonlinear system when the state variables depend only on time (finite dimension). It is done by transforming the system into a form of extended nonlinear normal form observability. The second one is constructed while considering the development of modelisation researches where engineers propose a new model of the biological system which state variables depend on both time and space (infinite dimension). This system model is called distributed parameter system. The conception of the observer is based on the backstepping technique which transforms the original system into an exponentially stable error system.