The move toward an ecosystem-based fisheries management (EBFM) requires new operational tools in order to support management decisions. Among them, ecosystem- and fisheries-based models are critical to quantitatively predict the consequences of future scenarios by integrating available knowledge about the ecosystem across different scales. Despite increasing development of these complex system models in the last decades, their operational use is still currently limited in Europe. Many guidelines are already available to help the development of complex system models for advice yet they are often ignored. We identified three main impediments to the use of complex system models for decision support: (1) their very complexity which is a source of uncertainty; (2) their lack of credibility, (3) and the challenge of communicating/transferring complex results to decision makers not accustomed to deal with multivariate uncertain results. In this paper, we illustrate these somehow theoretical “best practices” with tangible successful examples, which can help the transfer of complex system models from academic science to operational advice. We first focus on handling uncertainty by optimizing model complexity with regards to management objectives and technical issues. We then list up methods, such as transparent documentation and performance evaluation, to increase confidence in complex system models. Finally, we review how and where complex system models could fit within existing institutional and legal settings of the current European fisheries decision framework. We highlight where changes are required to allow for the operational use of complex system models. All methods and approaches proposed are illustrated with successful examples from fisheries science or other disciplines. This paper demonstrates that all relevant ingredients are readily available to make complex system models operational for advice.