The corrosion of steel bar is one of the main reasons for the lack of durability of concrete structure. The serious corrosion of steel bars can lead to mechanical strength loss of steel bars, cracking and spalling of concrete, eventually structural deterioration and failure, resulting in serious safety problems and huge economic losses. It is thus of great practical significance to control the corrosion of reinforced concrete structures for their long life and safe service. Steel bar corrosion inhibitor is one of the main measures used to slow down steel bar corrosion because of its advantages of simple construction, economical and effective. However, the problems of long corrosion test cycle and high cost become increasingly prominent. Amino acids, drugs, plant extracts, and green compounds are used as corrosion inhibitors with the continuous development of green organic corrosion inhibitors. Its composition is complex, the mechanism of action is complicated, and it is difficult to screen the effective ingredients. Molecular dynamics (MD) simulation is widely used in the research of corrosion inhibitors, which can explain the mechanism of corrosion inhibitors from the atomic level and help material control and test design. The effect and mechanism of corrosion inhibitor can be elucidated comprehensively via combining the methods of microstructure observation and substance composition tests. The MD simulation effectively solves the problem of long test cycle and high cost as one of the important methods to investigate corrosion inhibitors for steel bars, which has significant advantages and wide application prospects. This review briefly summarized the basic principles, common formulas and application processes of molecular dynamics simulation. The application of molecular dynamics simulation in mechanism research and property screening of steel bar corrosion inhibitors in recent years was represented. In addition, the research and application of MD simulation in corrosion inhibitors were prospected to provide a theoretical support for the development of simulation technology in corrosion inhibitors.The theoretical basis and common formulas of MD simulation are briefly introduced, as well as the environmental factors that need to be set in the application process and the basic simulation process are given. In the application of MD simulation, some appropriate parameters and conditions are set according to the research environment, and then the results can be calculated. The molecular adsorption and diffusion are used to determine the effect of corrosion inhibitors.The application of MD simulation in corrosion inhibitor research in recent years is summarized. In the application of corrosion inhibitor mechanism research, MD simulation is mainly used for a wide variety of organic corrosion inhibitors with a complex mechanism of action. It is reported that the corrosion inhibition mechanism of organic corrosion inhibitors is usually the formation of physical or chemical adsorption films on the surface of steel bars, and the adsorption behavior mainly depends on the physical and chemical properties of the corrosion inhibitor molecules. These properties are related to their functional groups, spatial structure and electron orbital properties, and the relevant calculation model is proposed. Based on the quantum chemistry calculation and molecular dynamics simulation, the adsorption energy of corrosion inhibitor molecules and steel bar surface can be calculated, and the most stable adsorption configuration can be simulated, so as to better explain the corrosion inhibition mechanism. The environmentally friendly, pollution-free and sustainable green corrosion inhibitors become a research hotspot, i.e., waste drugs, plant extracts, vitamins, DNA, green organic compounds, etc., as corrosion inhibitors. However, some of them are complex in structure and contain a variety of active ingredients. Conventional research methods need to carry out a large number of experiments, which consumes manpower and time. The MD simulation can be used to construct the motion behavior of different kinds of molecules on specific surfaces, and select the optimal molecules according to adsorption energy, diffusivity and barrier properties, thus improving the research efficiency of corrosion inhibitors.Summary and prospectsThe MD simulation method can obtain the thermodynamic structure and mechanical properties of complex molecules, analyze the dynamic evolution of the system, and obtain the time-dependent dynamic properties of the system. This can favor exploring the corrosion mechanism more conveniently and intuitively, and improving the research efficiency. The MD simulation becomes an effective tool to investigate the mechanism and development of steel bar corrosion inhibitors and has broad application prospects. However, the existing MD simulation cannot simulate complex chemical reactions, so it is more commonly used in the study of organic corrosion inhibitors, and the simulation method is relatively simple, mostly adsorption of an inhibitor molecule in pure water environment. Some studies show that a variety of ions in concrete environment can affect the effect of inhibitor molecules, and the inhibition effect is affected by molecular concentration and environmental temperature. In future studies, OH^- can be added to simulate the alkaline environment, and Ca²⁺, K⁺, Na⁺ and other cations can be introduced according to the real situation when simulating the concrete environment. The interaction between cations and inhibitor molecules can be intense in the simulation of phenolic and metal cation complexation reaction, thus obtaining more accurate and scientific simulation results. The use of MD to simulate the dynamic process of adsorption of multiple corrosion inhibitor molecules to film formation and corrosion barrier needs a further study. The study of the coupling adsorption of various inhibitor molecules on the surface of steel bars can favor expanding the application range of MD simulation and providing the data support for future prediction methods such as machine learning and deep learning.