Intense short-term wind events can flush multiple-inlet systems and even renew the water entirely. Nonetheless, little is known about the effect of wind variations at seasonal and interannual scales on the flushing of such systems. Here, we computed two Lagrangian transport time scales (LTTS), the residence and exposure times, for a multiple-inlet system (the Dutch Wadden Sea) over 36 years using a realistic numerical model simulation. Our results reveal pronounced seasonal and interannual variability in both system-wide LTTS. The seasonality of the LTTS is strongly anti-correlated to the wind energy from the prevailing directions, which are from the southwesterly quadrant and coincidentally aligned with the geographical orientation of the system. This wind energy, which is stronger in autumn-winter than in spring-summer, triggers strong flushing (and hence low values of the LTTS) during autumn-winter. The North Atlantic Oscillation (NAO) and the Scandinavia Pattern (SCAN) are shown to be the main drivers of interannual variability in the local wind and, ultimately, in both LTTS. However, this coupling is much more efficient during autumn-winter when these patterns show larger values and variations. During these seasons, a positive NAO and a negative SCAN induce stronger winds in the prevailing directions, enhancing the flushing efficiency of the system. The opposite happens during positive SCAN and negative NAO, when weaker flushing during autumn-winter is observed. Thus, large-scale atmospheric patterns strongly affect the interannual variability in flushing and are potential drivers of the long-term ecology and functioning of multiple-inlet systems.