Abstract: Drought is one of the most serious meteorological disasters in China and remote sensing technology shows greater potential for agricultural drought monitoring. Thermal inertia model derived from energy-balance equation has been proved to can monitor drought successfully in bare land and sparse vegetation coverage areas，but there are still certain limitations of the application in higher vegetation coverage regions. In this paper，MYD11A2 and MYD09A1 （DOY：081/113/115） derived from AQUA- MODIS （Moderate Resolution Imaging Spectroradiometer） data and relative soil moisture （RSM） in 10，20 and 50 cm soil depths near the surface were used to analyze the relationships between apparent thermal inertia （ATI） and RSM in March，April and May，2011. Besides，extended thermal inertia （ETI） established with vegetation index and temperature difference was applied to estimate surface soil moisture in higher vegetation coverage areas of Hebei Province. Results showed as follows：（1） Good correlation existed between ATI and RSM₁₀ near surface. Coefficients of determination （R²） of RSM₁₀-ATI fitting equations from March to May were 0.387，0.265 and 0.249，and the correlations between RSM₁₀ and ATI passed t-test at significance level α= 0.001. Both RSM₁₀ estimated from [RSM10-ATI] fitting equation and measured RSM₁₀ were analyzed and the mean relative error （MRE） values were 20.89%，28.91% and 31.54%，respectively. From March to May，with the vegetation coverage rising，the surface soil moisture estimation ability of ATI model was decreased. （2） ATI model in March was analyzed and result showed there was significant positive correlation between ATI and reciprocal of temperature difference. Hence，in high vegetation coverage area，thermal inertia model could be expressed as extended thermal inertia （ETI） coupled with vegetation index and reciprocal of temperature difference. In May 2011，area with ETI≤0 was 977.50 km²，which was mainly water body and distributed in the eastern part of Hebei Province. Area with 0<ETI≤0.015 was 112 140.05 km²，which mainly distributed in the northern，southern and southeastern parts. Area with 0.015<ETI≤0.030 was 58 513.31 km²，which mainly distributed in the central. Area with ETI>0.030 was 14 460.54 km²，which was mainly in the northeastern part. （3） In May 2011，drought monitoring by ETI model showed RSM₁₀-ETI fitting equation coefficient of determination （R²=0.359） was significantly higher than that of RSM₁₀-ATI（R²=0.249）. Moreover，MRE of RSM₁₀ estimated from ETI was 25.47%，which was markedly lower than that from ATI model （31.54%）. Furthermore，ETI was also used to estimate RSM in April to test its applicability and result showed that MRE of RSM₁₀ estimated from ETI model （25.47%） was lower than that from ATI model （28.91%）. So ETI model could retrieve surface soil moisture successfully in the area with higher vegetation coverage and the retrieval abilities of ETI for RSM₁₀ and RSM₂₀ were relatively higher than that for RSM₅₀.

Key words: high vegetation coverage, thermal inertia, drought monitoring