The nitrogen (N) balance (N input − N export by the grain) is often used to estimate the risk of N leaching from arable land. In a nine year study during the leaching seasons 1991/1992 to 1999/2000, the relationships between N fertilization, N balance and N leaching in the subsequent percolation period were investigated in a multifactorial field experiment near Kiel in NW Germany at the Hohenschulen Experimental Station. The crop rotation was oilseed rape (OSR) – winter wheat – winter barley, while soil tillage (minimum tillage without ploughing, conventional tillage), application of pig slurry (none, application in autumn, application in spring, application in autumn plus in spring), mineral N fertilization (none, 120 or 240 kg N ha−1 to cereals), and application of fungicides (none, intensive) were all varied. In each year the rotation and the treatments were located on the same plots. N leaching was calculated by multiplying the drainage volume with the respective N concentration obtained using ceramic suction cups. In all crops, N fertilization significantly increased the N balance in the order mineral N < spring slurry < autumn slurry. N leaching significantly correlated with the amount of percolation (P < 0.05). Highest annual N losses occurred with wheat following OSR (73 kg N ha−1) compared to the situation with OSR following barley (44 kg N ha−1). Without slurry, mineral N fertilization increased N leaching only slightly, whereas slurry, especially if applied in autumn, boosted N losses. Soil tillage and fungicide application had no significant effect on N leaching, although the latter showed a large influence on N balance. Increasing N balance progressively raised N leaching in the subsequent period with all crops, however, only 13–25% of the N balance surpluses originating from the preceding crop seem to leave the system via this pathway. The results indicate only a poor correlation between N balances and N leaching. In the short-term, therefore, the N balance is not an appropriate indicator for the environmental impact of N fertilization. However, if set up over a longer period, N balances may still give good estimates of the leaching potential arising from different management systems.