Multiband drought indices (DIs) have been developed for quantitatively characterizing the degree of region drought-hit and crop yield loss. Nowadays, drought indices are currently one of the hot topics with potential applications in agriculture, forestry and ecological environment monitoring. However, it is still lack of a systematic research for the multiband drought indices. This paper made a review of latest remote sensing derived drought indices. The concept of multi-dimensional spectral space was extended, where multi-dimensional variables are expanded from the remote sensing spectral bands to all the satellites derived products concerned with drought. Over 30 drought indices developed during the past years were categorized by constitution elements into four groups. Advantages and limitations of each group were analyzed. Further more, the family of drought indices was viewed as a “pyramid”, and the whole family’s composition were deconstructed from the bottom to the top. Environmental impacts like atmosphere, sensor calibration and soil effects on the drought monitoring are analyzed and expressed as improved factors into the models.