There exists a huge yield gap between given cultivar’s potential yield and on-farm experimental station’s yield aswell as actual yield on farmers’ field. It is essential to quantify the difference between these to devise strategies to fill thegap, and recommend the same to farmers to realize higher yield. This is very important especially for the crops grown underrainfed conditions as yield is often limited by water stress due to erratic rainfall patterns.Yield gap analysis studies are moreeffectively done using crop simulation models. A calibrated and validated DSSAT-CERES and DSSAT-CROPGRO modelswere used to run from 1985 to 2016 (32 years) for greengram (DGGV-2), soybean (JS 93), maize (BRMH-1) during kharifseason, and for sorghum (CSH-15R), chickpea (BGD-103) and wheat (DWR-162) during rabi season under potential(no water stress) and on-farm rainfed conditions on black clay soil in Northern Transition Zone (NTZ) of Karnataka.Results showed that during calibration and validation the model simulated the phenology, growth, yield and yield attributesof all the tested cultivars with high accuracy. The average of 32 years simulated outputs revealed that among the three kharifcrops the highest yield gap between potential and on-farm yields was noticed in maize (29.45 %) followed by soybean(16.45%) and greengram (14.97%). Among the three rabi crops tested the highest yield gap between potential and on-farmyields was noticed in chickpea (18.25%) followed by wheat (17.67%) and sorghum (10.28%). This study showed that evenunder NTZ where kharif season is believed to be more assured and more often than not enable take up double cropping,crops do experience moisture stress and give lower than potential yield. Thus needs strategies to narrow the gap.