Regression analysis is one method to determine the cause and effect relationship between one variable with other variables. Regression analysis is used extensively to make predictions, to know which predictor variables are related to response variables, and to know the forms of those relationships. If there are two response variables called bivariate regression analysis. In this study, because the data of response variables used distributed Gamma then the appropriate regression analysis is Gamma bivariate regression analysis (Bivariate Gamma Regression). In practice, it often occurs in the bivariate regression that the response variable is spatial data, example the measurement data containing a location information. The statistical method that has been developed by taking into account the spatial factor is Geographically Weighted Regression (GWR). Therefore, a statistical approach to spatial bivariate data will be developed with a continuous response variable that distributes Gamma bivariate through Geographically Weighted Bivariate Gamma Regression (GWBGR). Based on the result of the research, it can be concluded that the first derivative of GWBGR model is not closed form, it needs numerical analysis to get the parameter estimation value. The numerical analysis that can be used is Newton Raphson iteration. Testing the significance of GWBGR model parameters using Maximum Likelihood Ratio Test (MLRT) method.