In this seminar the status of satellite data assimilation is introduced for numerical weather prediction models at JMA, especially geostationary satellite data assimilation for cloud analysis, e.g. cloudiness and cloud water content, of Global Spectral Model (GSM). To calculate infrared brightness temperature and visible reflectance from GSM forecasts in all weather conditions a Radiative Transfer Model (RTM) had been developed based on the radiation process adopted in GSM. The brightness temperature and visible reflectance calculated by the RTM were compared with the observation of GMS-5 and GOES-9. Since the cloud distribution was similar to the observation, availability of the RTM for variational assimilation scheme was confirmed. The tangent liner and adjoint code are being developed for the variational scheme.