Soil hydraulic functions can be obtained with methods that range from complex and costly to simple and cheap. Decisions as to which is the most appropriate method for a specific application have to be based on a comparison of generated hydraulic functions. This comparison should preferably be based on a statistical comparison of practical applications calculated with the different hydraulic functions rather than on a statistical comparison of the functions themselves. In this study four different methods were used to generate hydraulic functions: (A) direct on-site measurement, (B) measurement in soil horizons in the area, ( C ) use of a national data set, and (D) use of Van Genuchten parameters correlated with soil texture and organic matter content. The four methods were compared by their effect on two practical aspects of soil behaviour: (1) evapotranspiration deficit and (2) flux through a plane at 30 cm below soil surface. These two aspects are highly relevant for agricultural and environmental use. However, direct measurement is not feasible. A validated simulation model was used for the calculations and results obtained with method A were taken as a reference. Calculations were performed for three soil profiles for a period of seven years. Deficits and fluxes, calculated with the four methods to generate hydraulic functions, were not significantly different using the data of the seven-year period. However, methods were significantly different when rainfall deficits were used as a covariable. This is true with the exception of downward fluxes in the period October until March which are most important for leaching of pollutants. The user has to decide whether differences between methods are sufficiently large to justify repeated, expensive on-site measurements (method A) or whether an investment will be made to make standard series of curves to be used everywhere (methods C and D).