Image classification is an important process of land use and land cover mapping. For effective image classification, quite a few aspects have to be considered including the accessibility of quality of satellite imagery, ground control points, a accurate classification method and the skill and proficiency of the user in the processes involved. This study classifies and maps the land use/land cover (LULC) of the Sagar Island using two images (1975 and 2015) and additionally verifies the precision of the classification method used. The study has been divided into two sections (1) Landuse/Land cover (LULC) classification and (2) accuracy assessment. Unsupervised classification was performed using Non-Parametric Rule and change detection was done for the 40 years study period. It was observed that 7.60% of mangrove vegetation’s were converted to cropland. Similarly, 40.26% of agricultural (mono-crop) land was converted to agriculture land, 1.48% of mud flat was converted to mangrove swamps, 1.87% wetland area converted to aquaculture land, and 22.54% agricultural (mono-crop) land converted to the settlement with homestead orchard respectively. Other LULC conversions are agricultural (mono-crop) land to cropland (40.26%), mud flat to shallow water (1.36%), wetlands to cropland (0.055%). The study had an overall classification accuracy of 79.53% and Kappa coefficient (K) 0.7465. This overall classification accuracy of the LULC maps is quite significant in terms of their potential use for land use change modeling of Sagar Island. © Springer Nature Singapore Pte Ltd. 2019.