Abstract: The applicability for monitoring and evaluation agricultural drought of five drought indices，including climate dry-wet index （I），relative moisture index （M），comprehensive meteorological drought index （Ci），precipitation anomaly percentage （Pa）and standardized precipitation index （SPI），was analyzed in Yunnan province by using of the correlation analysis and canonical correlation methods，based on the data of precipitation，average temperature，sunshine duration and soil moisture from August 2009 to July 2010 along with other climatic data in Yunnan The results showed that the explanation capability of these indices to the temporal-spatial variation of soil moisture （SM）was as follows：I＞M＞Ci＞Pa＞SPI，among which the explanation capability of the first three indices was 89.66%，85.47% and 82.59% respectively Pa and SPI were not able to express the changes of agricultural natural dry wet status accurately because of their primary expression for rainfall fluctuation Ci was still so sensitive to precipitation fluctuation that it was not suitable for being used in varied climates to a certain extent I and M distinguished natural dry wet changes well so were available to more immense areas They were superior to other three indexes while used in estimating agricultural drought and flood under various climates But both I and M had less susceptibility while used in the persistent severe-drought periods of long duration，when the length of duration time for index values below some critical values could indicate the intensity and persistence of drought

Key words: Soil relative moisture, Climate dry wet index, Relative moisture index, Comprehensive meteorological drought index；Agricultural drought, Applicability