The paper introduces FFA-Net, an end-to-end feature fusion attention network designed for single image dehazing. The network consists of three key components: a Feature Attention (FA) module, a basic block structure, and an Attention-based Feature Fusion (FFA) structure. The FA module combines Channel Attention and Pixel Attention to handle uneven haze distribution and different weighted information. The basic block structure integrates Local Residual Learning and Feature Attention, allowing the network to focus on more effective information. The FFA structure adaptively learns feature weights from the FA module, retaining shallow layer information and passing it to deeper layers. Experimental results show that FFA-Net outperforms previous state-of-the-art methods in both quantitative and qualitative metrics, achieving a PSNR of 36.39 dB on the SOTS indoor test dataset. The paper also includes a detailed introduction to image dehazing, related work, and implementation details, along with ablation studies to validate the effectiveness of each component.The paper introduces FFA-Net, an end-to-end feature fusion attention network designed for single image dehazing. The network consists of three key components: a Feature Attention (FA) module, a basic block structure, and an Attention-based Feature Fusion (FFA) structure. The FA module combines Channel Attention and Pixel Attention to handle uneven haze distribution and different weighted information. The basic block structure integrates Local Residual Learning and Feature Attention, allowing the network to focus on more effective information. The FFA structure adaptively learns feature weights from the FA module, retaining shallow layer information and passing it to deeper layers. Experimental results show that FFA-Net outperforms previous state-of-the-art methods in both quantitative and qualitative metrics, achieving a PSNR of 36.39 dB on the SOTS indoor test dataset. The paper also includes a detailed introduction to image dehazing, related work, and implementation details, along with ablation studies to validate the effectiveness of each component.