Purpose <p>Nitrogen (N) fertilisation is one of the main factors contributing to crop yield. Nevertheless, only a limited number of studies have addressed the consequences of spatial variability on the N balance (Nb). While the spatial variability of soil properties has been widely investigated, its influence on Nb has been analysed in only a few studies. Therefore, the objectives of this study were to compute a complete Nb over two growing seasons at various points in a field, and to investigate the relationship between Nb and soil properties.</p> Methods <p>To investigate the effect of soil properties on Nb, a linear multivariate regression (LMR) model, was compared with a geographically weighted regression (GWR) model, which evaluates spatial variability. The data were collected in Denmark over a field cropped with potato and barley for two years.</p> Results <p>The average Nb was − 127&#xa0;kg N ha^− 1 in potato and 65&#xa0;kg N ha^− 1 in barley, and its primary driver was crop N uptake. Clay, silt, and pH were the most important soil drivers in both models but their effect was highly dependent on the year and location. Overall, GWR outperformed LMR in terms of explained variability (84% versus 30%, on average) and root mean squared error (16 versus 34&#xa0;kg N ha^− 1, on average) in both years.</p> Conclusion <p>These results underline the importance of considering spatial variability when analysing N dynamics at the field level. Integrating the effect of soil properties on the N balance may promote more precise and sustainable fertilisation strategies.</p>