A feature of policy-making today is its dependence on scientific advice to deliver public policies that are robust, credible, and effective. This paper discusses how policy-making profits from scientific advice in areas where science and technology are significant. Particular attention is given to issues holding a high level of uncertainty, either because of inherent variability, because science is incomplete or controversial, or because data are inadequate to support a definitive answer. First, we analyse the social context that characterises the relationship between science and policy-making, with a focus on the decrease of public confidence in politicians and scientists. Second, we compare three different sets of guidelines on the collection and use of expertise in policy-making (issued by the UK, Canada and the European Commission, respectively) and identify two different approaches to scientific advice in policy-making. Third, based on a set of cross-national and multi-disciplinary case studies, we look at how the relationship between science and policy-making works in practice and propose a set of recommendations towards the establishment of a more robust and effective policy-making process.