Temporal mixture analysis (TMA) is deduced from spectral mixture analysis (SMA). They are algebraically identical except for that TMA is applied to temporal spectra and thus can extract the temporal characteristics of features. The ice concentration is diverse across the Antarctic sea through different periods, and TMA has a great potential to obtain this variability as an environmental normal. In the present study, sea ice concentration data remotely sensed by AMSR-E from 2003 to 2010 were used and seven typical endmembers were captured, standing for temporally different sea ice classification. TMA can also be utilized in change analysis of Antarctic sea ice concentration for its capability to record the spatial distribution of temporal characteristics, allowing further study of regional or global climatic variations. In short, TMA supplies a new method for researchers to investigate the spatial and temporal variability of polar sea ice.