Earth Radiation Budget Experiment (ERBE) data are used to validate radiative fluxes and cloud radiative forcing (CRF) simulated by the Laboratoire de M¿t¿orologie Dynamique (LMD) general circulation model (GCM). The emphasis of the work is on the development of new tests to obtain more significant elements of comparison between model simulations and satellite observations. These tests are applied to the clear-sky fluxes and the cloud radiative forcing. The validation of the CRF described by a model requires to test the consistency between the solar or shortwave (SW: 0.2 to 5 μm) and longwave (LW: 5 to 50 μm) cloud forcing. For this purpose, we compute the mean cloud perturbation of the planetary albedo as a function of the LW cloud forcing for both model results and ERBE observations. In the SW spectral domain, the consideration of total fluxes does not provide very constraining elements of validation because most of the observed variations are prescribed (incoming solar radiation, solar zenith angle). We therefore distinguish the part of the SW seasonal variations related only to the variation of external parameters (mainly the insolation) from the part which arises from the combined variation of internal climate parameters (mainly cloud albedo and snow/ice cover) with the insolation. Fourier analysis is used to study the seasonal amplitude and phase of the CRF. The seasonal variation of the cloudiness is, respectively, out of phase (in phase) with the insolation in mid-latitudes (in low and high latitudes). We show that this acts to enhance (to reduce) the seasonal amplitude of the absorbed SW flux in mid-latitudes (in low and high latitudes). Finally, we show that the impact of the seasonal variation of the cloudiness on the variation of the net CRF is less than 10 W m-2. ¿ American Geophysical Union 1992