Abstract: Composite materials are increasingly being used for critical components like aero-engine blades due to their excellent mechanical properties. Automated fiber placement (AFP) is essential for manufacturing these composite blades, but the path planning for complex blade structures poses challenges in terms of load capacity and stability. This study introduces a novel variable angle fiber laying path planning algorithm driven by adaptive principal stress, designed to address AFP path planning for large composite fan blades. Firstly, considering the complexity of the blade structure and the laying requirements, an adaptive principal stress-driven variable-angle placement path planning algorithm was proposed by mimicking biological structures found in nature. This algorithm can monitor and analyze the stress distribution at the current position in real-time and adaptively adjust the laying direction to better accommodate the structural characteristics and load-bearing performance of the blade. Secondly, an equidistant offset algorithm was employed to achieve uniform coverage of the blade surface, effectively balancing the overall trajectory's load-bearing capacity, layability and gap control quality. Lastly, through path simulation and detection of the curvature distribution, the engineering application feasibility and algorithmic viability of the filament winding trajectory were verified. The study demonstrates that this algorithm optimizes fiber laying paths, enhancing the efficiency and quality of path planning for variable angle laminate structures in fan blades, offering a valuable method for AFP in composite fan blade manufacturing.