Gas sweetening is a fundamental step in gas treatment processes. Aalkaloamines such as diethanolamine are one of the most generally accepted and wildly used solvents for gas sweetening process. Acid gas content in treated gas by alkanolamine solutions is one of the most important parameters for monitoring the performance of gas treating units therefore, it should be closely monitored to prevent operational problems in downstream processes and excessive energy consumption. In this study, a model based on an Artificial Neural Network (ANN) was used to predict sweet gas concentrations in absorption column. The model was applied to the data gathered from an experimental pilot plant. The experimental data includes DEA weight percent in lean amine, H2S, CO2 and H2O concentrations, temperature, pressure and flow rate of input and output gases and lean amine. According to the R value and mean square error of the network, ANN showed good accuracy of this type of modeling even for a wide range of variable parameters. So absorber column controlling can be simplified using the ANN Model and the process could be more efficient and economical.