Climate change will not only shift environmental means but will also increase the intensity of extreme events, exerting additional stress on ecosystems. While field observations on the ecological consequences of heat waves are emerging, experimental evidence is rare, and lacking at the community level. Using a novel "near-natural" outdoor mesocosms approach, this study tested whether marine summer heat waves have detrimental consequences for macrofauna of a temperate coastal community, and whether sequential heat waves provoke an increase or decrease of sensitivity to thermal stress. Three treatments were applied, defined and characterized through a statistical analysis of 15 years of temperature records from the experimental site: (1) no heat wave, (2) two heat waves in June and July followed by a summer heat wave in August and (3) the summer heat wave only. Overall, 50% of the species showed positive, negative or positive/negative responses in either abundance and/or biomass. We highlight four possible ways in which single species responded to either three subsequent heat waves or one summer heat wave: (1) absence of a response (tolerance, 50% of species), (2) negative accumulative effects by three subsequent heat waves (tellinid bivalve), (3) buffering by proceeding heat waves due to acclimation and/or shifts in phenology (spionid polychaete) and (4) an accumulative positive effect by subsequent heat waves (amphipod). The differential responses to single or sequential heat waves at the species level entailed shifts at the community level. Community-level differences between single and triple heat waves were more pronounced than those between regimes with vs. without heat waves. Detritivory was reduced by the single heat wave while suspension feeding was less common in the triple heat wave regime. Critical extreme events occur already today and will occur more frequently in a changing climate, thus, leading to detrimental impacts on coastal marine systems.