Background  The urgent need to respond to global climate change and to achieve carbon neutrality is driving geologic CO₂ sequestration technology to be increasingly large-scale, safe, and intelligent. Since the CO₂ sequestration process involves multi-physical field coupling, the engineering feasibility and long-term safety of CO₂ sequestration heavily rely on the capacity of numerical models to perform accurate characterization of complex subsurface processes. Therefore, establishing a model system, covering the entire process consisting of injection, migration, sequestration, and monitoring, that is suitable for various geobodies for CO₂ sequestration has emerged as a foundation for the engineering application of geologic CO₂ sequestration technology.

Advances  This study systematically expatiates on six types of core models for geologic CO₂ sequestration: multiphase flow, vertical integration, reactive transport, deep learning, CO₂ plume, and geomechanical models. Based on the practical validation through representative CO₂ sequestration projects across the world, this study develops a universal modeling methodology centered on the synergy between reservoir characteristics, algorithm selection, and monitoring requirements. Studies have revealed that in the injection phase, key parameters can be effectively optimized using Fourier algorithms coupled with the multiphase flow model; in the migration phase, the spatial distribution of CO₂ can be accurately traced using the plume model coupled with finite element and finite volume methods; in the monitoring phase, cap rock integrity and the fault reactivation risk can be systematically assessed using the geomechanical model, thereby enabling the full-chain dynamic safety characterization.

Prospects  Given complex geological conditions and the demand for long-term safe sequestration, future efforts should focus on intelligent modeling driven by both big data and artificial intelligence. It is advisable to establish a new generation of models that are capable of autonomous learning, real-time data assimilation, and dynamic optimization by deeply integrating geological mechanisms and multi-source monitoring data. The purpose is to significantly enhance prediction accuracy and scenario adaptability. This model system will further extend to a fully closed-loop system covering capture, transport, sequestration, utilization, and emission, thereby supporting the establishment and intelligent adjustment of the integrated scheme of the CO₂ sequestration-utilization cycle. Furthermore, this system will promote the transition of CO₂ sequestration from single-link simulation to whole-chain collaborative management. Overall, the results of this study provide a systematic methodology for the cross-scenario application of geologic CO₂ sequestration models and offer a pathway for the evolution of the model system toward prolonged effects and intelligence.