The Pluto-Charon system has been observed by the Imaging Photopolarimeter of the Infrared Space Observatory (ISOPHOT) in four filters at 60, 100 150 and 200 microns. In an attempt to determine the surface temperature distribution of Pluto, observations were performed in February, March and August 1997, covering in total 5 to 8 (depending on the filter) rotational phases of the Pluto-Charon system. The main results are the following: (i) the system is clearly detected at all four wavelengths (ii) the flux levels at 60 and 100 \mum are somewhat below those determined from IRAS (Sykes et al., Science, 237, 1336, 1987) but confirm that the brightness temperature is higher in the far-IR than at mm/submm wavelengths, (iii) the data indicate a rotational lightcurve, unambiguously at 60 \mum and more marginally at 100 \mum, providing a direct proof that Pluto's surface is not isothermal. With a maximum near an east longitude L=75\circ and a minimum near L=200\circ, this thermal lightcurve is roughly, but not exactly anticorrelated with Pluto's visible lightcurve, which has a mininum near L=100\circ and a maximum near L=220\circ (Buie et al. Icarus, 125, 233, 1997). In accordance with the visible lightcurve and HST imaging, the data can be modelled in terms of three different surface units (Charon, Pluto bright and cold" regions -- tentatively identified as N2 ice at ~ 40 K -- and Pluto dark and warm" regions). The dark and warm" regions are found to have a dayside temperature of at least 50-60 K and a thermal inertia of ~2\times104 erg cm-2 s-1/2 K-1. The low thermal inertia may contribute to the decrease of brightness temperatures towards longer wavelengths. Possible implications on the nature of Pluto's dark" terrains will be discussed.