Task of development the system for gas quality analysis based on articial neural networks have been studied. The structure of the proposed system is presented. The most important features and the main advantages of the proposed system are described. The main stages for neural network model development are shown. The architecture of the developed neural network model is described. Four metrics that are used for model accuracy estimation are provided. Recommendations are made for effective application of neural network tools as the main models for solving the task of the natural gas quality analysis.