<p>Hyperspectral images in practical applications require both detailed target contours and rich spectral information. Hyperspectral pansharpening technology, which is a fusion technique combining low-resolution hyperspectral images (LR-HSI) and panchromatic (PAN) images, can effectively address this dual requirement. Many existing pansharpening models still suffer from the imbalance between spectral fidelity and spatial detail enhancement caused by insufficient feature fusion in cross-modal feature interaction. To address this, this paper proposes an Adaptive Spatial-Spectral Cooperative Attention Network (ASSCAN) for hyperspectral image pansharpening. The Hierarchical Spectral Dynamic Calibration Module extracts multi-scale spectral features through a multi-layer local window partitioning and shifting mechanism. The Dynamic Contextual Spatial Attention Module captures long-range spatial contextual dependencies via a dynamic attention pooling mechanism to generate spatial attention weights. The Dynamic Cooperative Feature Fusion Module dynamically adjusts the contribution of dual-path features using learnable weights based on the outputs from the aforementioned two modules, achieving complementary and synergistic fusion of spatial-spectral features. Experimental results demonstrate that the overall performance of the proposed network surpasses that of the comparative methods across several evaluation metrics, and the effectiveness of each module has been validated through ablation studies.</p>