 The proposed AM-MFN Attention Mechanism Multiscale Feature Fusion Network is a novel approach to detect small objects in remote sensing images. It combines the attention mechanism and multi-scale feature fusion to improve the detection accuracy of small objects. The DHEM module enhances the characterization of small object features while the AMCC module reduces redundancy in the feature layer. Additionally, the NWD and GIOU are used to optimize the weights of the model for small objects and improve the accuracy of the regression boxes. Finally, an object detection layer is added to further improve the feature extraction ability at different scales. Experiments on two datasets demonstrate that the AM-MFN outperforms other state-of-the-art methods in terms of AP scores. This article was authored by Juensuo Chu, Zhongbing Tang, Lu Zheng, and others.