Space object detection is a key component of intelligent spacecraft operation systems. However, existing methods often face challenges such as insufficient feature extraction from objects and poor adaptability to exposure changes. To address these challenges, this paper introduces a space object detection algorithm based on multi-scale feature coupled attention (MSFC). First, the MSFC module is devised, combining channel attention with spatial attention mechanisms to thoroughly explore the correlations between multi-scale features, effectively enhancing the feature representation capabilities of objects in exposure variations. Second, the channel attention employs a transformer architecture to linearly map different channel features, dynamically adjusting the weight distribution of each channel to enhance the saliency of object features. The spatial attention globally models object spatial features through nonlinear mapping, further suppressing background interference, and improving the detection performance of objects under exposure variations. Final, the proposed algorithm is validated on the self-built dataset containing various exposure conditions and compared with State-of-the-art object detection algorithms. Experimental results demonstrate that compared to the baseline model YOLOv10n, the proposed method has improved \( mAP_{50} \) and \( mAP_{50-95} \) by 1.2% and 1.9% respectively, significantly improving detection accuracy and robustness to exposure variations while maintaining a lightweight model design.