Background - Evidence-based decision-making and practice are pivotal in public health. However, barriers do persist and they relate to evidence properties, organisations and contexts. To address these major knowledge transfer (KT) issues, we need to rethink how knowledge is produced and used, to enhance our understanding of decision-making processes, logics and mechanisms and to examine the ability of public health services to integrate research findings into their decisions and operations. This article presents a realist evaluation protocol to assess a KT scheme in prevention policy and practice at local level in France. Methods/design - This study is a comparative multiple case study, using a realist approach, to assess a KT scheme in regional health agencies (ARS) and regional non-profit organisations for health education and promotion (IREPS), by analysing the configurations contexts/mechanisms/outcomes of it. The KT scheme assessed is designed for the use of six reviews of systematic reviews concerning the following themes: nutrition, alcohol, tobacco smoking, physical activity, emotional and sexual life and psychosocial skills. It combines the following activities: supporting the access to and the adaptation of scientific and usable evidences; strengthening professionals' skills to analyse, adopt and use the evidences in the course of their practices and their decision-making process; facilitating the use of evidence in the organisations and processes. RAMESE II reporting standards for realist evaluations was used. Discussion - The aims of this study are to experiment and characterise the factors related to the scheme's ability to enable public health stakeholders to address the challenges of KT and to integrate scientific knowledge into policy and practice. We will use the realist approach in order to document the parameters of successful KT strategies in the specific contexts of preventive health services in France, while seeking to determine the transferability of such strategies.