针对城市低空空域因交通流量分布不均匀与建筑障碍密集而面临的安全管控难题，提出基于出行流量驱动的空域分层优化方法。该方法以空间分层规划为视角，通过优化各高度层的航向角范围，建立融合交通流特征与静态地理围栏约束的组合优化模型，以实现空域结构对动态运行需求的自适应调整。在求解策略上，设计了一种改进型遗传算法（FGA-ASO）。该算法通过混合初始化策略嵌入流量分布先验信息，采用流量加权交叉机制引导搜索聚焦于高冲突风险区域，并结合自适应变异策略平衡全局探索与局部寻优能力，从而实现空域分层结构的智能优化。最后，基于成都市OSM建筑数据与出租车OD数据构建仿真场景开展实验验证。结果表明，经FGA-ASO优化后的空域结构在安全性上显著提升，总冲突数降至386，较均匀分层基准降低26.9%，飞行器-建筑冲突数仅为1，平均侵入严重性下降7.0%。研究证明，以流量为导向的空域分层规划可有效利用垂直维度资源，显著增强城市低空运行的安全性，为空域精细化管理及城市空中交通系统的可持续发展提供理论依据与技术支持。

To address the safety management challenges arising from non-uniform traffic flow distribution and dense building obstacles in urban low-altitude airspace, this paper proposed a traffic flow-driven airspace layering optimization method. From the perspective of spatial layered planning, the heading angle range of each altitude layer was optimized. A combinatorial optimization model was established by integrating traffic flow characteristics and static geofencing constraints. This enabled the airspace structure to adaptively adjust to dynamic operational demands. In terms of the solution strategy, an improved genetic algorithm, namely the Flow-driven Genetic Algorithm for Airspace Stratified Optimization (FGA-ASO), was designed. A hybrid initialization strategy was used to embed prior traffic distribution information. A flow-weighted crossover mechanism was adopted to guide the search toward high-conflict risk areas. An adaptive mutation strategy was also incorporated to balance global exploration and local optimization capabilities. Consequently, the intelligent optimization of airspace layering was systematically achieved. Simulation experiments were conducted based on Open Street Map (OSM) building data and taxi origin-destination (OD) data of Chengdu. The results showed that the airspace structure optimized by FGA-ASO was significantly improved in safety. The total conflict count was reduced to 386, which is 26.9% lower than that of the uniform layering benchmark. The aircraft-building conflict count was only 1. The average intrusion severity decreased by 7.0%. This study demonstrates that flow-oriented airspace layered planning can effectively utilize vertical dimension resources. It significantly enhances the safety of urban low-altitude operations. The research provides theoretical support and technical pathways for refined airspace management and the sustainable development of urban air mobility systems.