Controlled drainage with subirrigation (CDSI) is a viable measure to supply, retain or discharge groundwater, thereby contributing to freshwater availability in agriculture under changing environmental conditions. Relatively simple CDSI systems can be controlled manually to set a few drainage levels. More advanced systems can be controlled remotely to set any drainage level (between a technical maximum and minimum). CDSI potentially improves hydrological conditions for crop growth, but the required external water supply can be large. Therefore, the objective of this paper is to investigate whether external water supply for subirrigation can be reduced by automatic control of CDSI systems in relation to crop water demand. Field measurements of a CDSI pilot in the Dutch sandy Pleistocene uplands were combined with weather forecasts to simulate the optimal drainage level and day by day water demand and supply using the agro-hydrological Soil, Water, Atmosphere, Plant model (SWAP). Firstly, model simulations showed that the water requirement reduced by 60 mm (dry growing season), 253 mm (average growing season) and 348 mm (wet growing season) using a dynamically managed crest level (CDSI-dyn) compared to using a fixed crest level (CDSI-fix), with minor effects on crop yield. Secondly, model simulations showed that a higher hydraulic resistance to downward seepage, a higher ditch water level or deeper roots reduced the water supply (up to 100 mm). Thirdly, accepting 10 % daily crop drought and oxygen stress for CDSI-dyn reduced the water supply requirement with 235–628 mm (dry vs wet growing season) compared to CDSI-fix. In conclusion, the required water volume for CDSI could be substantially reduced by automated control of the drainage level and water supply rate, while maintaining crop yield or accepting minor reductions, which increases the potential of implementation of CDSI systems.