Rock glaciers are often used as proxies for permafrost distribution in high-mountain environments where direct estimation of permafrost areas is challenging. These landforms are characterized by a mixture of rock and ice, and their movement and deformation provide important clues about the thermal state of the underlying permafrost. The presence of permafrost significantly impacts the stability of the landforms and the associated geomorphological processes, as well as the water balance in the surrounding catchment. Integrating different conditioning factors and rock glaciers makes it possible to infer the distribution and evolution of permafrost in these environments. In this study, we aim to use rock glaciers as a proxy for permafrost and to develop a hybrid permafrost model by combining both statistical and machine learning models. We have used the Frequency Ratio (FR) ensembled Artificial Neural Network (ANN) model to predict permafrost distribution over the Northern Kargil region of the Indian Himalayas. Out of 211 rock glaciers identified using highresolution imagery from Google Earth, 70 % (148) were used as a training dataset and the remaining 30 % (63) as a testing dataset. The study considered six influencing factors Slope, Aspect, Elevation, Mean Annual Air Temperature (MAAT), Solar radiation and Lithology for permafrost modelling. The results revealed that 30 % of the total geographic area has a high and very high probability of permafrost occurrence. The results have been validated by calculating the Area Under a Curve (AUC), which shows an accuracy of 96 %. The results of this study will contribute to the development of more cost and resource-effective approaches for probable permafrost mapping and improve the understanding of permafrost dynamics in changing climatic scenarios.