In this paper Normalized Difference Vegetation Index (NDVI) and Land Surface Temperature (Ts) were combined to indicate different land-cover types based on the fact that the biome has a similar seasonal trajectory in the NDVI/Ts space. Normalized Temperature-Vegetation Angel and Norm (NTVA &TVN) based on NDVI/Ts space, were put forward as input parameters for regional-scale land-cover classification. Remote sensing data used in this study are MODIS data products: MOD13 and MOD09, firstly the monthly Ts and NDVI were produced by the maximum value composite; secondly the monthly NDVI/Ts spaces were created; then NTVA &TVN were calculated for each of the 12 months. The monthly NTVA, TVN, NDVI, Ts were dealt with Principal Component Analysis (PCA) method, and their first three principal components were assembled to four groups as input parameters for classification. Remotely sensed land-cover system for China Based on land-ecosystem and maximum likelihood classifier were adopted to classify with four different input parameters. The classification accuracy for different inputs were compared and analyzed, and the results showed that combination of NDVI and Ts can indicate different land-cover types well; as input parameters, NTVA and TVN are applicable to macro land-cover classification, and can work well to improve classification accuracy at coarse spatial scales without other accessorial data.