Resumo. The objective of this work was to compare the indices of fire risk from hotspots and define the most efficient for the conditions of South Pantanal of Mato Grosso, based on meteorological data and number of hotspots. The work consisted of correlation analysis between some meteorological variables and the occurrence of hotspots as well as the comparison between the fire risk indices of Angstron (B), Logarithmic index of Telicyn (I), Nesterov Index (G), Monte Alegre Index (FMA) and Modified Monte Alegre Index (FMA+). Meteorological data were collected at 14 hours in the Nhumirim climatological station and hotspots during night of the NOAA12 and NOAA15 from 1999 to 2008. The wind speed and relative humidity were the meteorological variables which best correlated with the occurrence of fire. To detect any degree of fire risk, the FMA, the G and FMA+ were the most efficient indices. For the detection of degrees of fire risk classes very high and high, the G was the most efficient, followed by the FMA. Considering the high probability of success in detecting fire risks in any degree and also in the class of the most high risk, the FMA can be deemed more appropriate for estimating the fire risk for Nhecolândia sub-region.