In this paper we describe a method to map stochastic π-calculus processes in chemical ground form into hybrid automata. Hybrid automata are tools widely employed to model systems characterized by both discrete and continuous evolution and their use in the context of Systems Biology allows us to address rather fundamental issues. Specifically, the key ingredient we use in this work is the possibility granted by hybrid automata to implement a separation of control and molecular terms in biochemical systems. The computational counterpart of our analysis turns out to be related to the determination of conservation properties of the system.