Global climate change is accelerated by anthropogenic emission of greenhouse gases. Researches have shown that carbon (C) sequestration in agricultural soils can play an important role in off-setting industrial CO2 emissions. In this context, knowledge of soil organic carbon (SOC) stock at different scales is essential. Regional assessment of SOC stock is limited due to lack of adequate field observations which are cost and time prohibitive. The objective of this study is to quantify the SOC stock up to 1 m depth at the scale of Indiana State. We hypothesize that by capturing the variation in readily available secondary information that has effect on soil properties, we can predict the SOC content. A total of 524 soil profile data were used in this study which covers effectively the variation in landuse, weather data, and terrain attributes across the study area. Bulk density values which were not available at all the profiles were predicted using pedotransfer functions. Point observations of thirty years average temperature and precipitation data were interpolated to the state level using regression kriging approach. Depth distributions of SOC were modeled using negative exponential profile depth functions. Environmental parameters such as terrain attributes, weather data and landuse data were used to predict the parameters of the exponential functions using geographic weighted regression. The integral of the exponential function were used to predict the SOC stock. The SOC estimates will be validated using an independent validation dataset.