Abstract Simulations of bubble growth in porous media are carried out with a 2D numerical automaton built on a set of hypotheses derived from experimental observations at pore scale. Various types of 2D numerical networks are used as models of porous media with the aim to study the outcome of the combined effects of the wettabilitty (homogeneous and mixed) and the spatial correlation among the porous network entities on the evolution of a gas cluster growing by solute diffusion into it. The porous networks are constructed by using a variation of the dual site bond model (DSBM) framework. For the same pore-size distribution, the DSBM can provide simulated porous media of different topological structures. Under these scenarios, numerical simulations concerning the growth pattern and the growth rate of a single gas cluster are performed.