This chapter investigates and compares the interactions between science and policy (risk assessments and risk management) in five cases of environmental governance of the Baltic Sea: eutrophication, fisheries, invasive alien species, chemical pollution and oil discharges. An efficient interplay between science and policy is important for successful environmental governance, which applies particularly to the Baltic Sea where all five risks pose serious threats to environmental, social and economic aspects of sustainability. We use science-policy theory and an analytical framework based on a categorisation of relevant management responses linked to different states of incomplete knowledge (risk, uncertainty, ambiguity, ignorance) to investigate two main characteristics of science-policy interfaces: (1) organisational structures and (2) procedural aspects of managing scientific uncertainties and stakeholder disagreements. The analyses reveal differences and similarities in institutional and organisational designs of the respective assessment-management interactions, as well as in terms of how scientific uncertainties, stakeholder disagreements and socio-political ambiguities are addressed. All the five science-policy interfaces expose science-based management approaches that commonly are not able to cope sufficiently well with the complexities, uncertainties and ambiguities at hand. Based on our cross-case analyses, we conclude by recommending five key aspects that need to be addressed to improve science-policy interactions in Baltic Sea environmental governance: (1) more adaptive organisational structures in terms of time, context and place dependency, (2) increased knowledge integrations, (3) a more careful consideration of stakeholder participation and deliberation, (4) better management of uncertainty and disagreements and (5) increased transparency and reflection in the communication of science-policy processes.