The study of atmospheric concentration levels at a local scale is one of the most important topics in environmental sciences. Multivariate analysis, fuzzy logic, and neural networks have been introduced in forecasting procedures in order to elaborate operational techniques for level characterization of specific atmospheric pollutants at different spatial and temporal scales. Particularly, approaches based on artificial neural networks (ANNs) have been proposed and successfully applied for forecasting concentration levels of PM10, NO2, SO2, CO, and O3. The present study explores the development and application of ANN models for forecasting, 24 h ahead, not only the daily concentration levels of PM10 but also the number of hours exceeding the PM10 concentration threshold during the day in five different regions within the greater Athens area (GAA). The ANN modeling was based on measurements and estimates of the mean daily PM10 concentration, the maximum hourly NO2 concentration, air temperature, relative humidity, wind speed, and the mode daily value of wind direction from five different monitoring stations for the period 2001–2005. The evaluation of the model performance showed the risk of daily PM10 concentration levels exceeding certain thresholds as well as the duration of the exceedances can be successfully predicted. Despite the limitations of the model, the results indicate that ANNs, when adequately trained, have considerable potential to be used for 1 day ahead PM10 concentration forecasting and the duration within the GAA.