Glucocorticoids, including corticosterone (CORT), have been suggested to provide a physiological link between ecological conditions and fitness. Specifically, CORT, which is elevated in response to harsh conditions, is predicted to be correlated with reduced fitness. Yet, empirical studies show that CORT can be non-significantly, positively and negatively linked with fitness. Divergent environmental conditions between years or study systems may influence whether CORT is linked to fitness. To test this, we monitored free-living blue tits (Cyanistes caeruleus) during breeding over 3 years. We quantified foraging conditions during brood rearing, and examined whether they were correlated with parental baseline CORT and reproductive success. We then tested whether CORT predicted fitness. Elevated parental CORT was associated with lower temperatures, greater rainfall and lower territory-scale oak density. Whereas asynchrony with the caterpillar food peak was correlated with reduced nestling mass and fledging success, but not parental CORT. Only low temperatures were associated with both reduced nestling mass and elevated parental CORT. Despite this, parents with elevated CORT had lighter offspring in all years. Contrarily, in 2009 parental CORT was positively correlated with the number fledged. The absence of a direct link between the foraging conditions that reduce nestling quality and elevate parental CORT suggests that parental CORT may provide a holistic measure of conditions where parents are working harder to meet the demands of developing young. As the positive correlation between parental CORT and fledging success differed between years, this suggests that contrasting conditions between years can influence correlations between parental CORT and fitness. Ultimately, as CORT concentrations are intrinsically variable and linked to the prevalent conditions, studies that incorporate environmental harshness will improve our understanding of evolutionary endocrinology.