The complex flow field of the endwall and numerous design variables make it difficult to carry out effective design of film cooling hole arrangement. In order to achieve efficient design and enhance film cooling effectiveness, a Bayesian optimization-based approach for optimizing the arrangement of film cooling holes was proposed. Multiple parameters including hole location, compound angle, diameter, and spacing were optimized simultaneously. By integrating parametric modeling and surrogate model, the optimization process was conducted efficiently while significantly reducing simulation time and training costs. Compared to the baseline design, the film cooling effectiveness was improved by 68% and reduced total pressure loss by 10%. The results demonstrate simultaneous improvement in film cooling effectiveness and reduction in aerodynamic penalty, with an average 40% enhancement in effectiveness under variable mass flow ratio conditions. A design framework and a novel method are provided for optimizing film cooling and related geometric configurations.