This study is based on the visible, near infrared and thermal infrared channel data from theSsecond-generation polar-orbiting meteorological satelliteSof China (FY-3) equipped with the Medium Resolution Spectral Imager (MERSI). It firstlySidentified land surface ofSthe entire Chinese region in four times(2012.2, 2012.4, 2012.8 and 2012.11) by using Support Vector MachineSandSverifiedSclassificationSaccuracy. The results showed that: all the overallSaccuracy ofSclassificationSinSfour timesSare maintained above 95%. It proved that study based FY-3A/MERSI data usingSSupport Vector MachineSis highly accuracy and reasonable, and can be validly used in land surface type identification.