Parameters are determined for a model representing the drying dynamics of a wet fabric continuously passing through an infrared oven using a method involving off-line (batch) data processing, as well as by real-time recursive calculation. Drying trials with the infrared oven under various operational conditions demonstrate the strong nonlinearity of the drying process. Because of this, the model parameters established by off-line (batch) processing to not allow adequate calculation of the observed process responses under other different operating conditions. On the other hand, a method of recursive identification applied to realtime treatment of the readings from the oven provides a model with very satisfactory adaptation to changing operational conditions. The results also show that a simple first-order model represents the dynamic characteristics of the oven just as well as models of higher order, thus providing an important advantage for elaborating an efficient numerical controller for the drying process.