In Automated Fiber Placement (AFP), controlling defect propagation during layup is crucial for ensuring optimal part quality. Wrinkles are a common and complex AFP defect that, unlike gaps and overlaps, cannot be easily predicted based solely on basic geometrical conditions. Previous studies have modeled wrinkles using geodesic curvature and mechanical models, but these approaches were limited to single, well-defined surfaces. This paper presents a novel numerical approach for predicting wrinkle formation on generic tow surfaces during AFP. The methodology parameterizes tow surfaces into discrete segments and captures each top and bottom edge. By comparing these edges to an established critical steering radius, the model identifies critical regions which indicate potential wrinkle formation. Deformation geometry is then applied to each region using a tracing scheme that applies Rodrigues’ rotation matrix and an assumed cosine form. The efficiency and applicability of the proposed model is demonstrated through wrinkle predictions on both a flat plate and doubly curved surface. This methodology advances the understanding of AFP defect formation in process planning and offers a robust tool for optimizing layup strategies.