Abstract: Abstract: Agro-ecosystem, as the most active and controllable carbon pool in terrestrial ecosystem carbon cycle, can lead to substantial changes in the atmospheric CO2 concentration, thus affecting remarkably the global climate. The carbon cycle in agro-ecosystem is a complex process, which is influenced by factors such as climate, plants, soil properties and farm management. It is recognized that a model approach has an advantage in estimating spatiotemporal changes in carbon storage. Carbon cycle models are considered to be the most effective means to study carbon cycle. This paper emphasized on the carbon cycle process of agro-ecosystem, introduced the transference and the mechanism of carbon cycle between different carbon pools, identified characteristics of different models in association with carbon cycle of agro-ecosystem from 1960s, summarized and analyzed the application of international carbon models and others invented and developed in China in the agro-ecosystem. During these years, several models, such as RothC, CENTURY, DNDC, EPIC and APSIM have been widely used to estimate carbon changes at national or global scales. These models provide understanding of carbon flow through food webs and explore the role of carbon storage in the whole agro-ecosystem. They also allow analysis of environmental risks and provide a guide to know more about the relationship among carbon, nitrogen and water cycle. More recently, some new carbon models have been developed in China for simulating the carbon budget of agro-ecosystems. For example, the Agro-C can simulate crop net primary production via Crop-C sub model and changes in soil organic carbon via Soil-C sub model under various conditions of climate, soil, and agricultural practices, which makes it possible to extrapolate the model to a wider domain. Validation of the Soil-C sub model suggested that an inappropriate simplification of the carbon flow between various C pools may introduce errors into the estimates. Carbon loss in croplands was a very important issue for Chinese agro-ecosystems, and there was still a great potential for substantially improving carbon status by adopting proper farming practices and land-use pattern. This review has drawn together various perspectives on some of the key issues that should be addressed in order to keep more carbon and enhance soil productivity of croplands. In particular: 1) the mechanism of carbon cycle should be explored to model changes in carbon for the whole agro-ecosystem; 2) the study spatiotemporal changes of the carbon cycle should be investigated further; 3) the relationship among the nitrogen cycle, the water cycle, microbial and the carbon cycle should be made clear; 4) to use carbon cycle model to account for a range of different management practices and the redistribution of carbon; and 5) to investigate potential effects of different management practices on the carbon distribution for the government to formulate relevant policies.