Early forecasting of pollen concentration in the atmosphere is very important for medical applications due to the increasing occurrence of allergic diseases induced by allergenic pollen. Moreover, flowering and pollen dispersion are of great interest for agronomic studies of plant productions. Several statistical techniques have been used to forecast pollen concentration in the air and concern the prediction of the start of the pollen season, the maximum airborne pollen concentration, and the date when this occurs. Some forecasting techniques are based on the analysis of airborne pollen time series and, in addition, of meteorological variables involved in the phenomena. A time series is a set of measurements of a variable taken over time at equally spaced time intervals. The most frequently used time series models include the autoregressive integrated moving average (ARIMA) models; recently, artificial neural networks (ANN) have been applied in time series modelling and forecasting, due to their good performances with complex and non-linear phenomena. The aims of this study are (I) to develop both the ARIMA and neural network models, (II) to analyse and compare the performance of these models and (III) to improve the accuracy of airborne pollen forecasting for the principal allergenic plants of the Mediterranean area. The study was carried out on aerobiological and meteorological data collected from 1986 to 2000 in the urban area of Sassari, Italy; the sampling device was a Burkard volumetric spore trap. The meteorological data collected by a weather station of the Sardinian Agrometeorological Service were air temperature, air relative humidity, wind speed and direction, and rain intensity. The data set was divided into two sections: the first section, composed of eleven years of data was used to calibrate the models, the second, composed of four years, was used to test and validate the models. The analysis was performed on Graminaceae and Oleaceae pollen. For each species, Autocorrelation Function (ACF) and Partial Autocorrelation Function (PACF) were used to calculate the significant autocorrelation existing in the airborne pollen data and to identify the components of the ARIMA models. ANN models were realised using a three-layer feed-forward topology and the backpropagation learning optimisation algorithm. These ANN use as input the calibration airborne pollen time series; in addition, another ANN model was realised in order to forecast the airborne pollen time series using the values of air temperature, air relative humidity and rainfall. The ARIMA(1,1,1)365 was used to forecast the daily airborne pollen concentration: the error of the date of maximum concentration range from seven to twelve days on Graminaceae pollen. The model ARIMA(1,1,1)365 performed less well on Oleaceae (12-15 days). ANN models perform much better and the error ranged from six to eight days on Graminaceae; good results were also obtained with the alternative approach, based on the development of an ANN using as input the meteorological variables involved in the phenomena. The study verifies the capabilities of ANN as a tool for forecasting some characteristics of the pollen season that can be used to support preventive allergic therapy.