<p>In sheet metal forming processing, the drawbead plays an important role in manufacturing feasibility. However, the design of optimal Drawbead Restraining Force (DBRF) remains a challenging problem. This paper introduces an innovative methodology for the optimization of DBRF design that enhances both efficiency and accuracy. This approach involves a hybrid Physical-Artificial Intelligence (Physical-AI) method to establish an initial design that aligns closely with the acceptable range. Initially, analysis in mechanics of metal forming is undertaken to discern part features correlated with the DBRF. Subsequently, multiple AI methods are deployed to create an initial DBRF design, utilizing features gleaned from pre-existing designs of analogous parts. The DBRF design is then modelled as an optimization problem. To streamline this process, we introduce the Nearest Point Adjustment (NPA) method, designed to narrow the variable domain for ensuing iterations. By determining dominant bead sections in relation to the location of critical strains, the NPA method facilitates the identification of the optimal DBRF design with fewer iterations. The developed methods were applied to the DBRF design of two automobile panels, hood inner and outer, and the results reveal significant improvements in both computational efficiency and accuracy.</p>