Abstract: Remote sensing change detection technology aims to observe and analyze remote sensing images of the same area at different times to determine the locations, types, and extent of changes. Addressing the issues of blurred edges between changed cropland and background and boundary adhesion between changed cropland in existing methods, a cropland change detection network based on difference enhancement and edge awareness (DGEANet) is proposed. DGEANet uses a Siamese ResNet to extract multi-scale, multi-level features from bi-temporal remote sensing images. The proposed Cross Attention Difference Enhancement module receives multi-scale bi-temporal features, enhancing the difference and correlation between temporal features. Edge-aware module and edge-guided module are introduced at each scale to identify cropland boundaries, improving change detection accuracy while enhancing edge integrity of the changed areas. Additionally, a parallel attention fusion module calculates spatial-channel attention in parallel and performs pixel-level detail recovery, enabling effective extraction of changed cropland of varying sizes. Auxiliary edge-aware loss is used to apply deep supervision to change boundaries, improving network training and enhancing the integrity of extracted change regions. DGEANet achieves high detection accuracy on standard cropland change detection datasets, CLCD and PX-CLCD, demonstrating the method's effectiveness.