From 2004 to 2013, Korea has experienced a total of 174 flood disasters and has a total estimated cost of USD 7.32 billion. However, reports showed that the total expenditure of the government amounted to 1.4 times the estimated losses and damages and the private companies have spent twice the said estimated amount. To summarize, the post-disaster loss and damage reports showed underestimated values. In this regard, the National Emergency Management Agency (NEMA), the government institution designated to assess and analyze the damages and losses as well as evaluate the disaster risks of the said areas in accordance to their disaster risk management plans, are now developing a new estimating method for damages and losses. This study aims to develop flood damage functions that will estimate the flood damages of Gunsan City based on the building type: residential, commercial and agricultural facilities, by utilizing the Ordinary Least Squares Regression and later on, the Geographically Weighted Regression. The model building process includes flood depth, flood duration, inundated area, family income and land price as the parameter variables. Due to normality issue, the datasets were transformed through Box-Cox Method. Both Ordinary Least Squares (OLS) Regression and Geographically Weighted Regression (GWR) were evaluated in this study, but the search for ‘best fit’ resulted to the use of GWR.