In recent years, significant changes in precipitation regimes have been observed and these manifest in socio economic and ecological problems especially in regions with increased vulnerability such as the Mediterranean region. For this reason, it is necessary to estimate the future projected precipitation on short and long-term basis by analyzing long time series of observed station data. In this study, an effort was made in order to forecast the monthly maximum, minimum, mean and cumulative precipitation totals within a period of the next four consecutive months, using Artificial Neural Networks (ANNs). The precipitation datasets concern monthly totals recorded at four meteorological stations (Alexandroupolis, Thessaloniki, Athens, and Patras), in Greece. For the evaluation of the results and the ability of the developed prognostic models, appropriate statistical indexes such as the coefficient of determination (R 2 ), the index of agreement (IA) and and P. Ralli Str., 122 44 Athens, Greece 1980 K.P. Moustris et al.the root mean square error (RMSE) were used. The findings from this analysis showed that the ANN's methodology provides satisfactory precipitation totals in four consecutive months and these results are better results, than those obtained using classical statistical methods. A fairly good consistency between the observed and the predicted precipitation totals at a statistical significance level of p < 0.01 for the most of the examined cases has been revealed. More specifically, the Index of Agreement (IA) ranges between 0.523 and 0.867 and the coefficient of determination (R 2 ) ranges between 0.141 and 0.603. The most accurate forecasts concern the mean monthly and the cumulative precipitation for the next four consecutive months.