Policy makers and farmers use tools, such as a nutrient balance, to gain insight into the environmental impact of agricultural practices. A discrepancy, however, exists between the needs of policy makers and farmers, about the use and the spatial scale of such tools. Farm balances calculate nutrient balances across all agricultural fields within a farm without distinguishing separate fields, whereas field balances calculate a nutrient balance on a delineated field. For farmers, a nutrient balance at field level is more useful than at crop or farm level, because decision making and fine-tuning management occurs at the field level. A field balance, however, requires more detailed data than a farm balance and therefore is less easy to implement. As soil types influence nutrient balances, we hypothesize that if within-farm variation in soil types is low, there is no need to replace a farm balance by a field balance. To test this hypothesis, we computed nutrient balances at farm and field level on five Dutch dairy farms (three on sand, two on clay), varying in degree of within-farm variation in soil series. A full year of soil nitrogen (N) and phosphorus (P) input and output data on farm and field level were provided by farmers, while soil variation was determined using the Dutch 1:50.000 soil map. The Annual farm Nutrient Cycle Assessment (ANCA) was used to calculate soil N and P surpluses, and soil nutrient fluxes such as nitrate leaching and nitrous oxide emission at farm and field level. Even on farms with few soil series, a considerable variation in N and P inputs, outputs and balances across fields was found, due to management differences and soil properties not represented by the soil map. Furthermore, field-level balances better represented nitrogen leaching than farm-level balances on farms with diverse soils (reflected by different leaching factors) and negative nitrogen field balances (deficits). Also, using field balances, for one case study farm the highest soil N surplus (kg ha−1) was found on grass fields with the highest risk of N leaching. A field balance, therefore, provides more meaningful information than a farm balance when variation in soil types and/or management factors is found within the farm, because soil types and management factors affect N and P balances, N leaching and N emissions. For farms with the highest variation in soil types and/or management, we recommend using field-level nutrient balances in order to detect extreme surpluses, deficits, leaching and/or emissions, to improve management decisions.