Crop phenology is essential information when evaluating crop production in the food insecure regions of West Africa. The only currently available global product that includes phenological variables is the MODIS Land Cover Dynamics Yearly (MCD12Q2) product. This product is produced each year at 500 m spatial resolution, from the 8-day vegetation index EVI (Enhanced Vegetation Index) calculated from the NBAR reflectance (Nadir Bidirectional Reflectance Distribution Function - Adjusted Reflectance). In order to analyze the information content of MODIS MCD12Q2 product, the phenological variables were extracted for areas previously identified as cropped on eight specific sites in Mali and compared to rainfall data and expert knowledge on Malian agriculture. MODIS MCD12Q2 data analysis showed that only 70% of the cropped pixels in Southern Mali had a complete phenology information on the whole vegetation cycle (four phenometrics values), and that a large part of the pixels displayed unrealistic late Start-Of-Season (SOS) values. A close analysis of the original EVI data indicated that these inconsistent SOS values were due to missing EVI data during the vegetation development phase (due to the presence of cloud cover) conducting to a false detection of SOS. We then proposed a simple way to correct the SOS values. In Africa, food security systems could benefit from such a phenology product, by utilizing its spatially continuous information in agrometeorological modeling, and thus improving agricultural production estimation.