Considering the increasing uptake of the concept of "ecosystem services" in landscape management and environmental policies, it is urgent to establish a consensual framework to assess the complex relationships among ecosystem services, considering both the supply- and the demand-sides. A diversity of approaches have been proposed to evaluate ecosystem services associations, but not all methods are equivalent and methodological choices need to be made depending on the scientific and policy questions at hand, as well as the type of data available. Based on previous classifications of ecosystem service associations, we propose to characterize three broad types of associations considering the ecological (supply side) and socio-economical (demand side) aspects of ecosystem services: supply-supply, supply-demand and demand-demand. We then review quantitative methods available and propose guidelines to assess those three categories of relationships among ecosystem services and identify their explanatory variables following three steps: (i) detecting ecosystem services associations, (ii) defining bundles and (iii) identifying the explanatory variables of ecosystem services associations. For each step, strengths and weaknesses of different statistical analysis and machine learning methods are described. The proposed interdisciplinary methodological approach takes one step toward embracing such complexity of socio-ecological systems as it considers ecosystem services delivery (supply-supply), stakeholders' needs (demand-demand), and on how stakeholders can benefit from the ecosystem services delivery (supply-demand). We illustrate how such a diverse spectrum of methods may apply for land management.