Habitat destruction is one of the main threats to environmental integrity. Assessing the consequences of human impacts is crucial both to predict and prevent structural and functional changes of habitats. However, to date almost all studies on marine threats, from regional to global scales, have been entirely qualitative and generally based on little more than expert opinion. We have developed a meta-analytical approach to quantify overall effects of various stressors on different Mediterranean habitat types and to compare the relative importance of different impacts across a range of habitats. We first qualitatively reviewed and synthesized 366 experiments (either manipulative or correlative) collected in the literature. After a selection procedure, we finally quantitatively meta-analyzed 158 experiments. We showed that fisheries (destructive or not), species invasion, aquaculture, sedimentation increase, water degradation and urbanization have negative effects on Mediterranean habitats and associated species assemblages. We also explored the overlap between the impacts identified as important in the Mediterranean and those identified by experts as being important globally, highlighting the inadequacies of relying on expert opinion alone. Finally, we drew attention to the critical lack of empirical knowledge about marine systems in many areas of the Mediterranean, which impedes the implementation of effective conservation measures. Our study is the first to synthesize experimental analyses on human-driven impacts on marine habitats across such a broad geographic scale.