IRAS data have shown that active K and M emission line dwarfs, the
 so-called flare stars, are brigther at IR wavelength than non emission
 line dwarfs of the same spectral type. The infrared excess, first
 appearing at wavelengths as short as 12 micron, becomes more prominent
 at 25 and 60 micron. Several mechanisms and models can be invoked to
 explain the origin of the IR excess: (i) syncrotron radiation of
 relativistic electrons produced during flare and microflare activity;
 (ii) emission from a stellar wind favoured by the presence of a hot
 corona; (iii) circumstellar dust surroundings flare stars consistently
 with their young ages (post T-Tauri phase) or originating from mass
 ejection episodes that are associated with strong flares.
 In order to choose the best model among the above possibilities, the
 reconstruction of the energy distribution from optical to radio
 spectral domains is vital and has already been done for a few sample
 objects, but using a limited spectral coverage.
 However, only a complete description of the far IR spectral region
 will allow us to discriminate among the above hypotheses, because only
 in the far IR region the slope of the spectrum changes significantly
 depending on the physical mechanism responsible for the emission.
 The improved sensitivity of ISO, with respect to IRAS, will allow us
 to obtain much better values of the IR fluxes, and, more important,
 beyond the present 100 micron limit. At present, for most dMe sources,
 only upper limits are known, within a typical 2 sigma error.