Accurate segmentation of polyps in colonoscopy images is of vital importance for the early diagnosis and treatment of colorectal cancer (CRC). However, accurate segmentation is a challenge due to the diversity of polyps in size and shape as well as unclear boundaries. To this end, we propose a novel edge-guided network (EGNet) for polyp segmentation, which achieves cross-level feature fusion by effectively utilizing edge information and enhances the focus on polyp edges. Specifically, we first propose a multi-scale enhancement module (MSEM), which strengthens the feature representation capability through the interaction between convolutional kernels of different sizes. Next, we propose an edge-aware guided module (EAGM), which extracts more discriminative edge features and enhances the model’s sensitivity to edge details. Finally, we propose a cross-level fusion module (CLFM) that integrates contextual cues from different levels to enhance the semantic perception of target regions in complex backgrounds. Our experimental results on four benchmark datasets show that EGNet outperforms other state-of-the-art methods.