Increasing soil carbon content is an environmentally friendly measure to mitigate global warming. Mycorrhizal fungi have been regarded as an important part of soil carbon pool, and at least 78% of land plants form symbiotic relationships with arbuscular mycorrhizal fungi (AMF). Therefore, it is of great significance to study the soil organic carbon (SOC) sequestration potential via the arbuscular mycorrhiza pathway for improving ecosystem carbon sinks. However, there is currently a lack of systematic exploration of the SOC sequestration mechanism via AMF pathway. AMF exhibit distinctive ecological characteristics, including a higher turnover rate than roots, extensive hyphal expansion, facilitation of mineral-organic matter complexes, and transformation of metabolic products into stable soil carbon components. These features collectively contribute to the AMF-mediated carbon sequestration mechanism. Specifically, AMF acquire carbon from plants and convert it into AMF-derived carbon through the growth of hyphae, metabolic products (particularly glomalin-related soil proteins), and residuals. Furthermore, AMF extraradical hyphae coexist and collaborate with other microbes, enhancing the decomposition of plant litter, promoting microbial anabolism and turnover. This process not only increases the input of plant-derived carbon and the accumulation of microbial-derived carbon but also promotes the formation of soil aggregates, effectively protecting soil carbon from decomposition and thereby enhancing soil carbon accumulation via the AMF pathway. Soil carbon storage through the AMF pathway is most significant in forests, followed by grasslands, and least in croplands, closely associated with climate change, biotic and abiotic soil factors, the underground shared mycorrhizal network, and human activities. This review also explores the differences in soil organic carbon (SOC) storage via the AMF pathway across various ecosystems, underscoring the current gaps in research and outlining prospective areas for future investigation. For instance, while plants are known to allocate photosynthates to AMF, the spatiotemporal dynamics and peak values of carbon availability to AMF across different climate zones or ecosystems remain poorly understood. Additionally, the distribution characteristics of SOC sequestered through the AMF pathway in various soil layers, along with the direct, indirect, and cumulative contributions of AMF to the SOC sink across ecosystems, require further clarification. There is also a need for systematic research into how the structure of plant communities influences plant carbon inputs through the AMF pathway and the responses of microbial communities. Furthermore, it is crucial to analyze the dynamics of the 'trade-off’ between plant carbon inputs and the stabilization of existing soil carbon. These areas represent critical avenues for enhancing our understanding of the ecological roles of AMF in carbon sequestration processes. This article aims to better understand the mechanism of SOC sequestration via the AMF pathway, providing theoretical reference for research on enhancing ecosystem carbon sinks through the mycorrhizal pathway.