Abstract Understanding how canopy-scale photosynthesis responds to temperature is of paramount importance for realistic prediction of the likely impact of climate change on forest growth. The effects of temperature on leaf-scale photosynthesis have been extensively documented but data demonstrating the temperature response of canopy-scale photosynthesis are relatively rare, and the mechanisms that determine the response are not well quantified. Here, we compared leaf- and canopy-scale photosynthesis responses to temperature measured in a whole-tree chamber experiment and tested mechanisms that could explain the difference between leaf and crown scale temperature optima for photosynthesis. We hypothesized that (i) there is a large contribution of non-light saturated leaves to total crown photosynthesis, (ii) photosynthetic component processes vary vertically through the canopy following the gradient in incident light and (iii) seasonal temperature acclimation of photosynthetic biochemistry has a significant role in determining the overall temperature response of canopy photosynthesis. We tested these hypotheses using three models of canopy radiation interception and photosynthesis parameterized with leaf-level physiological data and estimates of canopy leaf area. Our results identified the influence of non-light saturated leaves as a key determinant of the lower temperature optimum of canopy photosynthesis, which reduced the temperature optimum of canopy photosynthesis by 6–8 °C compared with the leaf scale. Further, we demonstrate the importance of accounting for within-canopy variation and seasonal temperature acclimation of photosynthetic biochemistry in determining the magnitude of canopy photosynthesis. Overall, our study identifies key processes that need to be incorporated in terrestrial biosphere models to accurately predict temperature responses of whole-tree photosynthesis.