Abstract: Artificial neural networks for the Osprey process based on back-propagation are established and some tests are conducted. The relationship between the processing parameters and the outputs are simulated with the networks. It is shown that neural networks are successfully used to pridict porosity values of the process.