From the Voyager 1 mission in 1980 and from gound-based observations to date, Titan's atmosphere has been revealed to us in terms of temperature and composition. The main consituent is N2, followed by CH4 and H2. N2+CH4 photochemistry leads to the formation of other gas molecules found in trace amounts (hydrocarbons, nitriles and oxygen compounds). CO2 was discovered in the infrared Voyager data with a stratospheric abundance of the order of 10-8 - 10-9. CO was detected from ground-based near-infrared spectroscopy in 1986 with a tropospheric mole fraction of about 5 x 10-5. In 1988 a stratospheric mixing ratio as low as 4 x 10-6 was derived from millimeter heterodyne observations leading to the suggestion that CO may be depleted in the stratosphere with respect to the troposphere, the transition occuring somewhere between 20 and 60 km. This depletion is not expected due to the very long photochemical lifetime of CO. The fates of CO and CO2 are tightly linked according to photochemical models. To reproduce the CO2 Voyager abundance, a CO mole fraction of 1.1x10-4 is required, marginally consistent with the near-IR value in the troposphere, but disagreeing with the millimeter CO abundance. Are the CO2 or the CO measurements wrong ? To discriminate between the two, the vertical profiles of CO and CO2 need to be determined with high precision. CO presents no lines in the thermal IR range and its rotational lines in the submillimeter range are too weak to be detectable by ISO/LWS. On the other hand, ISO/SWS Fabry-Perot mode offers the opportunity to observe the CO2 emission band centered at 667 cm-1 with great precision: in this range, the spectral resolution achieved by ISO is about 215 times higher than Voyager and permits to resolve the band, thus yielding information on the CO2 abundance at different atmospheric levels. The CO2 vertical profile will in turn constrain the photochemical models from which the CO stratospheric abundance can then be derived thus solving the problem of its distribution in Titan.