Distances to subdwarfs and absolute luminosities of those stars have been one of the poorest known astronomical data for many years. Thanks to the observations made from Space by HIPPARCOS accurate distances to 15 to 20 subdwarfs will be finally available. In order to take full advantage of these new data, for which one of us is PI, with early dat access in two HIPPARCOS programmes,it is necessary to use these distances to derive the basic structural parameters ,Radii,Effective Temperatures, and Bolometric Luminosities of the objects. That would be easy, if the angular diameter theta was known, through the relationships: 2R=d*theta sigma*Teff**4=f*(theta/2)**(-2) Where R is the radius of the star, d its distance, got from HIPPARCOS, sigma Stefan's constant,and f the integrated flux received on earth. But theta is much too small to be accurately measurable by interferometry for a dwarf star at several tens of parsecs. Therefore we propose to use the so called Infrared Flux Method (IRFM) which derives theta from: fnu = Fnu*(theta/2)**2 = Fnu*(R/d)**2 (1) where fnu is the monochromatic flux in the infrared received from the star at the earth, and Fnu the monochromatic flux emitted at the surface of the star.Integration of (1) over nu gives then both the effective temperature and the Luminosity.Fnu is derived by fitting a theoretical model representing the atmosphere, using a large number of observables (profiles of the Balmer lines,ionization equilibria, colours,etc... ) and fnu is observed. The method has been applied with broad bands (H,J,K of Johnson) instead of spectrophotometric fluxes, but discrepant results are found from the different bands, difficult to synthesize from model atmospheres, mostly because of uncertainties in the filter characteristiscs and conversion to absolute photometry. We propose to use the spectrophotometric capability of ISO (PHT40) complemented by PHT05 for obtaining actual monochromatic fluxes directly comparable to model atmospheres.