In light of the projected climate change for the coming decades, there is an urgent need for multifunctional landscapes that are capable to provide a diversity of ecosystem services. This requires a better understanding of social and ecological factors that influence how these landscapes are managed and how this, in turn, influences the provision of ecosystem services. Land Use Land Cover (LULC) changes are one of the main factors that lead to spatiotemporal changes of ecosystems services. As such, the identification of the main socioeconomic drivers of LULC can give important insights about the drivers of ecosystem services. However, the analysis of ecosystem services in a context of socio-ecological systems is still underdeveloped. Brazil has witnessed intense changes in LULC in the last five centuries, which may have influenced the provision of ecosystem services at local, regional and global scales. In the southeast mountain area of the Atlantic Forest biome, the Zona da Mata de Minas Gerais is characterized by a heterogeneous landscape mosaic composed of pasture and coffee fields intermingled with forest fragments, which are predominantly inhabited and managed by family farmers. The Zona da Mata is considered a complex socio-ecological system and is an interesting case to study the spatio-temporal provision of ecosystem services. In Chapter 2, I assessed the LULC changes from 1986 to 2015 and their main socioeconomic drivers. By combining data obtained from satellite images, workshops and secondary data, I showed that forest and coffee areas increased, and pasture decreased. These changes were associated with government measures to protect the environment, financial support of family farmers, migration to cities and the agroecological movement. A scenarios analysis of contrasting socio-economic narratives indicated that sustainable measures taken by the government to protect the environment and support family farmers with financial credit will lead to increase forest and coffee areas in the Green Road scenario. In contrast, the socioeconomic development in the Fossil Fuel scenario, which projects a decline in environmental protection and focuses on rapid economic development, there will be a decline in forest areas. In Chapter 3, I explored the spatial variation of ecosystem services from 1986 to 2015 and the impacts of LULC changes on ecosystem services provision levels and their interactions. To map the spatio variation of ecosystem services, I used the LULC maps from 1986 and 2015 (Chapter 2) and the InVEST model. This analysis indicated that the conversion of forest to pasture has strong negative impacts on soil erosion control and water flow regulation, manifesting mostly as trade-offs and dis-synergies between ecosystem services. In Chapter 4, I investigated the separate effects of LULC changes and climate on water dynamics from 1990 to 2015, and explored scenarios of LULC change and climate change for 2045. For this purpose, I used the SWAT model and climate data combined with historical and future LULC maps developed in Chapter 2. I found that the variation in climate variables was the main factor for the observed increase in the river streamflow in the study period and that forest can buffer extreme precipitation events. The exploration of future scenarios indicated that the increase in forest cover under the Green Road scenario is expected to decrease the surface runoff water and increase evapotranspiration as compared to the Fossil Fuel scenario, mitigating the impacts of soil erosion and climatic extremes in the region. Projected changes in precipitation and temperature are expected to have negative impacts for agriculture in the future. In Chapter 5, I assessed the impact of climate change on the suitability of Coffea arabica production in the study region and the potential of agroforestry systems to mitigate these impacts. For this, I combined the species distribution model MaxEnt with current and future climate projections. Agroforestry system have the potential to reduce air temperatures under the canopy of trees. I explored the effect of the altered the microclimate in agroforestry systems on the suitability for coffee production by adjusting future climate data to reflect conditions in agroforestry systems. I found that the area suitability for coffee production from the current monoculture coffee systems will decline by 60% under the projected climatic changes. However, the implementation of coffee agroforestry systems can mitigate these negative impacts of climatic change and maintain 75% of the area suitable for coffee production in 2050. Combining social and ecological systems in an interdisciplinary framework, generated insights in the relationships between climate and LULC change, and how this influences several ecosystem services. This framework connects different research fields and allows different stakeholders to work together to find effective ways to work towards multifunctional landscapes that promote the sustainable use of ecosystem services.