Nineveh province in Iraq has experienced a process of land cover conversion and vegetation loss especially in last decades. It is important to get accurate information on vegetation loss and changes in areas that are used for agriculture.Among the most effective methods to study and get information about this phenomenon is remote sensing technology.
Since classical approaches lack of accuracy, artificial intelligence has been introduced to strengthen feature detection which leads to better classification.This paper uses ant colony algorithm to study and classifypart of Nineveh province land into sixclasses. These are Agriculture land/flood plane, Water, Outcrop, Origin of early sand sheet, Desertable area, and Sand dunes.The variation in these six classes from 1987 to 2009 is shown. Results show that agriculture region and flood plain decreased from around 31% in 1987 to 11.2% of total area in 2009 while origin of early sand sheet and desertable area increased from 42.7% to around 49%. Beside that sand dune appears in 2009 to form about 26.47% of total area under study.

Keyword-Remote Sensing, Classification, Ant colony algorithm, Vegetation Loss.