Variability in the feeding ecology of young fishes over short and long time scales in estuaries is likely to affect population dynamics. We studied 14 years of early stage Striped Bass feeding ecology in the Hudson River Estuary over a 25-year time span, including years in which invasive zebra mussels markedly altered energy flow within the estuary. We predicted that feeding success would be low and that diet composition would be altered during years of high zebra mussel impact, particularly in upriver locations where mussels occur. Feeding success in the short term was indicated by volume of gut contents and in the long term by dry mass at length, i.e. condition; these measures were positively intercorrelated and varied significantly year to year. We tested for associations between condition and multiple biotic and abiotic environmental variables. There was a strong negative effect of zebra mussel grazing rate on condition in upriver locations and a weak positive effect in downriver locations. In upriver locations, condition was 33% higher when local salinity was high and zebra mussel grazing rates were low, whereas in downriver locations, condition was 35% higher when zebra mussel grazing rates and copepod abundance were high and local dissolved oxygen was low. Copepods, amphipods, mysids, and Leptodora constituted the highest prey-specific index of relative importance throughout the estuary. There was no evident effect of the zebra mussel invasion on diet composition. This long-term study corroborates the inferences of earlier studies that zebra mussels reduced early-stage striped bass growth rate.