The EUROS model, a numerical grid model using simple linear chemistry, is applied routinely to predict winter smog episodes in the Netherlands. The required meteorological input consists of prognostic fields obtained from KNMI or ECMWF. The model calculated hourly concentration fields of SO2, SO4, NO2, NOx and NO3 for an area covering practically the whole of Europe. The effective prognosis period is +72 hours. From an operational-technical point of view, the model appears to function excellently ; it is only in exceptional cases that model results are not available before 9.00 am. It is not yet possible to provide a more detailed evaluation of the system ; on the one hand since the system has not yet been operational throughout a complete winter period, and on the other since there have been hardly any episodes with raised pollutant concentrations. Superficial analysis reveals that the model can lead to both over- and underestimates of concentration levels. Reasons for discrepancies between model predictions and measurement results were usually to be found in the (automatically generated) input parameters. Recommendations are made to improve the quality of the model's predictions.