Under-fertilisation with potassium (K) in potatoes can lead to significant yield loss, while over-fertilisation can reduce the starch content. Further, the cost of using fertiliser unnecessarily reduces the economic payoff. Thus, it is relevant to apply the correct amount of K fertiliser based on the soil K status, which can vary with-in the field. A model for redistribution in the field is available in CropManager; a software programme for precision farming, from which allocation maps can be sent directly to farm machinery. The model is based on soil K status, soil type, predicted yield, date of soil sample, correction for type of fertiliser and finally, pre-vious L-application in the season is subtracted, so only the remaining K-need is displayed. The model ena-bles a more even coverage of K-requirements within the field and thus a more uniform crop. According to the model, more clayey soils have a lower K-requirement than more sandy soils. Soil samples taken close to the time of fertilisation will give a more correct graduation model and preferably 1 sample per ha.