Abstract: A two-dimensional cellular automaton (CA) model has been developed for the description of the normal grain coarsening process. The probabilistic CA method incorporating Moore's definition of the neighbourhood is used to simulate the normal grain coarsening process with a new transition rule. The model simulates the grain coarsening process in as much detail that is possible, from the point of initial nucleation to subsequent coarsening with computational times. The unique result is that the grain coarsening speed can be controlled by the specific method, this result is vital to model the grain coarsening quantitatively. In order to make this model valid, experimental work has been done to provide solid evidence for this model. Comparison of the modelling result and the experimental result has been carried out.