This study aimed to develop an algorithm to retrieve surface soil moisture in Taiwan from the moderate resolution imaging spectroradiometer (MODIS) data (1-km resolution) and AMSR-E (advanced microwave scanning radiometer-earth observing system on-board aqua satellite) soil moisture data (25-km resolution). Data were processed for 2009 comprising three main steps: (1) computing the temperature vegetation dryness index (TVDI) by empirical analysis of the relationship between land surface temperature (LST) and normalized difference vegetation index (NDVI) data. Because the climate was altered by elevation, the study area was divided into two regions (using a threshold of 1000 m) based on the digital elevation model (DEM) analysis. The TVDI was thus processed separately for each region, (2) converting TVDI to the same unit with AMSR-E soil moisture data (mg cm-3) by regression analysis of the two datasets, and (3) validating the results of retrieved soil moisture. The retrieved soil moisture showed the comparable spatiotemporal patterns with the AMSR-E soil moisture data. The goodness of the methods was assessed by comparing the TVDI results with the retrieved soil moisture using root mean square error (RMSE) indicated that the methods in which the study area was partitioned into two regions produced slightly better results (RMSE = 36.7 mg cm-3) than the methods in which the study area was not partitioned into two regions (RMSE = 39.8 mg cm-3).