Estimating reservoir pressure and temperature is one of the challenges of geothermal reservoir engineering. A new proxy model has been developed to estimate reservoir pressure and temperature using wellhead pressure, temperature and non-condensable gas (NCG) amount. An exhaustive set of wellhead data covering a range of possible wellhead pressure, temperature and NCG data is used in a calibrated wellbore simulator, which is then used to create a knowledgebase to train an artificial neural network (ANN) model. The developed ANN model tested using numerical simulation model results coupled with a wellbore simulator. It is observed that the ANN model can accurately estimate reservoir pressure and temperature for various production scenarios in liquid phase reservoir.