Process modelling represents an important area of research and application in aluminium industry. Many physically-based models have been proposed and validated. However, advanced statistical methods based on neural networks represent an alternative when more conventional models are non-existent or unsatisfactory. Two of these methods, Gaussian Process modelling and dynamic neural networks, have been examined on the test problem of the ageing of aluminium-copper alloys. Both methods can predict ageing curves after training on experimental data. They could therefore potentially be developed further and applied to more complex processing problems.