Lane detection plays an important role in the autonomous driving system and attracts widespread interest in recent years. Despite the advantages of conventional work, such as segmentation-based and point-based methods, these models either practically require a large sum of anchors or only leverage the connection message of discrete points in the lane. Different from this, curve-based methods enable one to naturally learn the holistic representation of lanes based on the geometric semantics of curves and conduct end-to-end optimization conveniently, thus serving a promising direction. However, the existing curve-based (e.g., polynomial-based and B´ezier-based) methods encounter draws in: 1) modeling some complicated curves mathematically. 2) Simultaneously capturing geometric semantics within and between lane curves. 3) The endpoints of lane lines are prone to occlusion, which is critical for curve modeling. To tackle these issues, we revisit the curve-based methods and propose a novel model Dynamic NURBS Network (DBNet). Specifically, we further introduce the NURBS curve to model lanes, enabling to theoretically fit various complicated curves and guarantee the robustness of local and global optimization. Based on this, we introduce three key modules to address lane detection challenges. Firstly, we propose a local dynamic-interaction module that adaptively exploits the geometric message of local structure. Secondly, a global association-sharing module to capture the global features of lanes, which can together promote the semantic association and tackle the occlusion problem inside and outside the lane. Lastly, a curve fitting enhancement module to enhance the feature of both ends of the lane lines. Moreover, the proposed method achieves a new state-of-the-art performance on three public benchmarks and especially achieves 77.98% of the F1 score on curve scenario in the CULane dataset.