Ecological network analysis (ENA) is an efficient tool to conduct holistic evaluation of the functioning and structure of ecosystems. ENA results can be used to support management decisions, but further development is needed to improve ENA application. We compared the food web functioning of three areas of the Wadden Sea: the Sylt-Rømø Bight, the Norderaue tidal basin and the Jade Bay, and present an example application of ENA indices in decision making processes. We used a sequential increase in uniform uncertainty from 50% to 99% using 10% increments to evaluate the robustness of the network comparisons. The results showed that the Jade Bay differed in its system functioning from the Sylt-Rømø Bight and the Norderaue tidal basin which represent the Northern Wadden Sea. The Jade Bay system, which is dominated by mudflats, had a heavier reliance on detritivory pathways, while the Northern Wadden Sea, which is dominated by Arenicola-sand flats, relied more on exchanges with the North Sea. Higher recycling, redundant pathways and flow diversity in the Jade Bay indicated that this system is probably more resilient against external perturbations than the Northern Wadden Sea systems which are more organized and specialized. This was related to the high standing stocks of suspension feeders in the Northern Wadden Sea, resulting from the establishment of several invasive suspension feeders, such as the Pacific oysters, the American razor clams and the common slipper shell. This study showed that 1- uncertainty analyses are crucial for the interpretation of ENA results and their use in management, 2- temporal measurable trends of indices which are robust to model uncertainty would be more appropriate for decision making than single reference values, 3- using ENA for assessment purposes in the Wadden Sea must be based on several representative sites taking into account their habitat types as well as morphological and physical characteristics, in terms of water depth, hydrodynamics and degree of enclosure of back barrier areas.