The modelling of chemical and biochemical systems is highly dependent on the reaction network of the system. A reaction network should contain all the reactions which can occur in addition to the species involved. However in many cases there will be reactions occurring which are missing from the network. These are called gaps (Reed et al., 2003; Duarte et al, 2004) and occur because the databases of reactions used to build them are incomplete. The main aim of this work is to identify all the reactions which can occur in a given system including the unknown ones. This is accomplished through a new procedure which will be called metabolic network development and is aimed at constructing reaction networks for biochemical systems.