Abstract: As an important fossil fuel, natural gas plays an important role in the global energy structure and energy transformation. The natural gas treatment technology is significant for improving product gas quality, ensuring safe operation of equipment, and meeting environmental regulatory requirements. However, due to the complex coupling of multi-phase state, mass transfer, heat transfer, and multivalent state chemical reactions in the production process, it shows strong nonlinearity and dynamic uncertainty, which makes the traditional modeling method based on mechanism model limited in accuracy and computational efficiency. In recent years, data-driven modeling techniques based on a machine learning algorithms has demonstrated high flexibility and excellent predictive accuracy by revealing intrinsic patterns in process data. However, due to the absence of mechanism constraints, its interpretability and extrapolation are still obviously insufficient. In order to cope with the above challenges, the hybrid modeling technology combining mechanism and data-driven has gradually become a research hotspot. Hybrid modeling technology fully integrates the physical and chemical interpretability of the mechanism model and the flexibility and efficiency of the data-driven model, which can improve the accuracy, computational efficiency and robustness of modeling in complex chemical engineering processes modeling. This paper provides a comprehensive review of the research progress for modeling technologies based on mechanism modeling, data-driven modeling, as well as hybrid modeling, and discussed their applications prospects and challenges in natural gas treatment process modeling.