Abstract: The steelmaking process needs to complete a series of operations such as decarburization, desulfurization, dephosphorization and deoxygenation of steel, which is a complex multifactorial control process, with high and inhomogeneous temperatures in the reactor, simultaneous occurrence of multiphase chemical reactions, and mutual coupling between phases in terms of mass, momentum and heat transfer. The accurate prediction and control of the steelmaking process has always been a difficult problem in iron and steel smelting and a hotspot for metallurgical workers. The advance in recent years on the reaction thermodynamics and kinetics of steelmaking process and its application in numerical simulation were summarized in the current study. In terms of reaction thermodynamics, the main models for the activity calculation of the liquid steel are Wagner Interaction Parameter Formalism (WIPF), Unified Interaction Parameter Formalism (UIPF), and the Associate Model. At the present stage, the WIPF model is still the most widely used model for liquid steel activity, but with the development of new steel grades, the universality of this model has encountered challenges. There is an urgent need to develop a new model for liquid steel activity calculation and to supplement it with new data. The main models for slag activity calculations are Molecular Theory, Ionic Theory, Regular Ionic Solution Model, Modified Quasi-chemical Model, and Ion and Molecular Coexistence Theory. Reaction kinetic models such as the Multicomponent Coupled Reaction model, the Effective Equilibrium Reaction Zone model, and the unreacted nucleus model can accurately predict the changes of the molten steel, slag, and non-metallic inclusions over time in the steelmaking process. However, the mass transfer coefficients in the kinetic models were mostly determined by empirical equations, which cannot accurately characterize the kinetics in different reactors. By coupling the reaction kinetics with three-dimensional numerical simulation, the three-dimensional distribution of the molten steel composition and its evolution with time during the iron desulfurization process in KR and the steel decarburization process in RH have been revealed. However, there is no mature three-dimensional numerical simulation for the multi-phase and multi-dimensional reactions of the integrated consideration of molten steel, slag, inclusions, refractory materials, alloys, etc., which requires further in-depth study.