Social-ecological system (SES) simulation model can be used as an effective tool for integrating landscape patterns and policy research in interdisciplinary fields, and it also can address complex feedback mechanism between social-ecological effects of landscape patterns and relevant landscape-based policy decision. The research progress of SES models was analyzed in the field of forest-agriculture landscape patterns, using bibliometric analysis and literature review (including 3091 papers from dataset of Web of Science) for identifying integrated research trend of the social-ecological methods from 2000 to 2019. The main conclusions are as follow: 1) A conceptual framework was developed for integrating forest-agriculture landscape patterns and social-ecological processes. Using this framework, we found that most of existing models in cases were insufficient to analyze the complex feedback relationship between landscape processes and socioeconomic decision, and the major type in this feedbacks was one-way impacts of land use on its ecological consequences. 2) A comparison of typical categories in social-ecological models was conducted by using literature review, based on dimensions of feedbacks, landscape, spatiotemporal scale, uncertainty and complexity. We address that there is an urgent need to develop diverse methods of scenario-based simulation and integrated landscape effects, for improving theoretical and methodological knowledge of 4 SES models, including empirical statistical, system simulation, pattern optimization, and agent-based model (ABM). Specially, the ABM model have better performance to reflecting socio-ecological feedback for its capability of bottom-up modeling and making-decision functions. 3) We provided a framework for elaborating complex conceptual links between potential pathways in social-ecological system models. Base on the framework, we suggested that the SES models can be promoted by several available pathways, such as, developing scenario-based simulation, improving the theoretical fundaments of decision making in changes of landscape patterns, strengthening model uncertainty analysis, reducing model complexity and integrating qualitative and quantitative data. Especially, we addressed that the integration of pattern optimization model and agent-based model should be given to a priority level for development of integrated modeling research, as both models can be used to synthesize the characterization of major socio-ecosystem processes by using spatiotemporal and explicitly spatially methods. The results of this study is useful for generating a comprehensive understanding for the role of forest-agriculture landscape pattern in complex social-ecological systems at multi-scales. Furtherly, the inductive discussion and conclusions can be used to support the development and application of integrated social-ecological models in related emerging interdisciplinary fields, such as sustainability science, ecological economics and land system science.