Photosynthetic phenotypic parameters of winter wheat are important indicators of canopy physiological status, photosynthetic function, and crop growth. However, canopy-scale hyperspectral estimation of these parameters remains affected by canopy structural heterogeneity, environmental variation, and mixed spectral signals. This study evaluated the contribution of visible (VIS), near-infrared (NIR), and shortwave infrared (SWIR) regions and their combinations to estimating photosynthetic phenotypic parameters of winter wheat. Field experiments were conducted under three nitrogen application levels and 65 winter wheat genotypes, and a total of 507 valid canopy-level samples were used for model development and validation. Competitive adaptive reweighted sampling (CARS) was used to select characteristic wavelengths, and partial least squares regression (PLSR), Bayesian ridge regression (BR), and backpropagation neural network (BPNN) were applied to construct estimation models. Model performance was assessed using R², RMSE, and RPD. Results showed that NIR-based models achieved the best overall performance, with the highest validation R² of 0.828 for photosynthetic rate. The VIS + NIR combination showed stable predictive ability across multiple parameters, whereas SWIR-only models showed limited performance, with R² values below 0.5 for most parameters. Photosynthetic rate, intercellular CO₂ concentration, performance index on an absorption basis, and chlorophyll a content were predicted more accurately than the other traits. These findings indicate that canopy hyperspectral data can support quantitative monitoring of photosynthetic phenotypic parameters, and that NIR-related structural and scattering information plays a key role in winter wheat canopy phenotyping.