When a biological system is modeled using a mathematical process, the following step is normally to estimate the system parameters. Despite the numerous computational and statistical techniques, estimating parameters for complex systems can be a difficult task. As a result, one can think of revealing parameter-independent dynamical properties of a system. More precisely, rather than estimating parameters, one can focus on the underlying structure of a biochemical system to derive the qualitative behavior of the associated mathematical process. In this talk, we will discuss study of qualitative behaviors of stochastic biochemical systems using reaction networks. A reaction network is a graphical configuration of a biochemical system. One of the key problems in this field is to relate dynamical properties and the underlying reaction network structure. The goal of this talk is to 1. walk you through the basic modeling aspect of the stochastically modeled reaction networks, and 2. to show how to derive stability (ergodicity) of the associated stochastic process solely based on the underlying network structure.