This paper assesses the relationship between gridded (5 × 5 °) monthly North Atlantic (10-70 °N, 80 °W -20 °E) seasurface temperature anomalies (SSTAs) and concurrent, one-monthly and two-monthly lagged rainfall totals for four coherent Icelandic precipitation regions over the period 1961-2002. Monthly correlation fields are derived and tested for field-significance at the 0.1 level. Stepwise multiple linear regression models are then constructed to predict one-monthly (t + 1) and two-monthly (t + 2) lagged rainfall from SSTA variations for those months and regions with field-significant correlation patterns. Most of the field-significant correlation patterns are for concurrent relationships and for precipitation during the winter months in northern Iceland at the t + 1 and t + 2 timesteps. The t + 1 and t + 2 predictor patterns of northern Icelandic precipitation in February are field-significant. For this month, above-average precipitation in northern Iceland in February is linked to a pool of warm water that tracks eastwards from the eastern Atlantic (40-55 °N, 5-40 °W) in December into northern European sea areas (e.g. North Sea, Norwegian Sea) by January. Study results also reveal that December precipitation in northern Iceland is also potentially predictable from SE North Atlantic SST (10-40 °N, 5-45 °W) in the preceding October and November. The models' skill is assessed retrospectively by re-deriving the forecasting algorithms over training periods A (1961-1990) and B (1971-2002), and then evaluating the models' performance over validation periods A (1991-2002) and B (1961-1970). Coefficients of determination (R 2 ) between observed and modelled t + 1 and t + 2 February and December precipitation in northern Iceland range from 4 to 49%, indicating some utility of these forecasting models for seasonal predictions of Icelandic precipitation.