Sounding retrieval algorithm has been developed for processing the Geosynchronous Imaging Fourier Transform Spectrometer (GIFTS) clear sky radiances. The algorithm is a statistical retrieval followed by a nonlinear iterative solution. In the statisitical retrieval procedure, the training data set of global radiosonde profiles used for establishing the synthetic regression algorithm was enhanced through the addition of more radiosonde observations, particularly in under-represented weather regimes and the polar regions. IR surface emissivity measurements sorted by ecosystem type will be used to improve the surface characterization in the training data, and a more physical basis for assigning a skin temperature to each training profile has been investigated. Simulation shows GIFTS is able to capture both the spatial and temporal features of atmospheric soundings. MODIS and AIRS data on EOS’ Aqua satellite are also used for testing the algorithm for GIFTS sounding retrievals.