Nowadays, satellite rainfall estimation data are used in climate impact studies due to the low density of the rain gauge network in sub-Saharan Africa and the difficulties in accessing data from national meteorological services. However, limitations may exist in the effectiveness of these micro-scale data. Therefore, studies to evaluate the performance of these satellite products are always necessary before their use. This work aims to assess the validity of the rainfall estimation data from «African Rainfall Climatology – version 2» (ARC-2) in the cotton production zone of Mali from 1983 to 2002. The rainfall estimation data were extracted for 10 weather stations and compared to rainfall measurements from these weather stations. Pearson correlation test was used to check the relationship between the two datasets for each station. The comparison of daily rainfall values shows ARC-2 estimated rainfall values very close to the measured rainfall values at the stations, even if the ARC-2 data were found to be sometimes overestimated. The ARC-2 rainfall data also closely follows the annual cycle observed for monthly and annually measured data. The precipitation correlation coefficient between the ARC-2 data and the measured rainfall data is approximately 0.70, with a significance level of 99% at the monthly scale. After calibration, precipitation measured by ARC-2 reproduces almost 99% of the observed data, thus confirming that ARC-2 data can be used for climate studies in Mali.