Abstract: Hyperspectral images are used to characterize the objects with unprecedented accuracy of the data. The active learning aims at providing efficient training set by iterating the samples. This study reviews the concepts involved in active learning algorithm for classification of remote sensing image or hyperspectral image. The diversified vision of hyperspectral sensors was awakened with the latest development of remote sensing and geographical information. Imaging spectroscopy which is commonly known as hyperspectral remote sensing was recently inspected by researchers and scientists for exploring vegetations, minerals, etc. This hyperspectral imaging requires large data sets and new processing techniques. Several active learning algorithms are implemented in hyperspectral images for better classification and greater accuracy.