Land use and land cover dynamics are a result of the interactions between human activities and the environment. The objective of this thesis is to analyze Amazonian land use and land cover pattern dynamics in order to identify the underlying system dynamics. By combining empirical statistical models and Fuzzy Cognitive Mapping, feedbacks in the human-environment system can be explored to identify more sustainable development pathways. The results show that specific feedback loops can lead to a sustainable human-environment system in the Brazilian Amazon, e.g., in case policies such as Payment for Ecosystems Services (PES) and Reducing Emissions from Deforestation and Forest Degradation (REDD) are enforced. Also, the analysis indicates that land market regulations and the enforcement of the Forestry Code can reduce deforestation. It is concluded that policy effectiveness of sustainable land use practices can be better evaluated by using the combination of statistical and cognitive methods. In summary, the thesis illustrates that added value in analyzing land system changes is achieved if insights obtained at different scales are combined through different methods of analysis.