This thesis presents an implementation method for optimizing the external geometric dimensions of an existing wireless power transfer (WPT) coil through multi-objective optimization. Wireless charging systems have been widely applied in daily electrical devices, and the trade-off between the geometric dimensions of the charging system, its charging efficiency, and power transfer capability is a key challenge faced by designers and manufacturers. During the design process, the evaluation methods for the power, losses, weight, and size of wireless charging coils significantly influence the product design cycle as well as the labor and time expenses associated with the design process. Based on an existing WPT coil sketch, this thesis designs and verifies the feasibility of implementing a multi-objective optimization method. The proposed optimization method is developed based on given power transfer requirements and external dimension constraints. A 3D geometric model is reconstructed using the SALOME open-source modeling platform, where meshing is performed to prepare the geometry for finite element analysis (FEA). The ElmerFEM open-source finite element solver is then employed to evaluate the coil's performance from multiple perspectives. To achieve large-scale iterative optimization for a single performance evaluation, a multi-objective constrained optimization framework is formulated in a Python environment, where constraint equations are defined and deployed using Pymoo. To address the challenges encountered in the implementation of this design method, this thesis primarily considers two key aspects. The first is the automated performance evaluation of a given coil geometry using Python. Given a set of geometric parameters, SALOME can be automated using Python scripts to generate the geometric model and perform mesh generation through a descriptive approach. The generated mesh files are then processed by the ElmerFEM solver, which, supported by SIF configuration files, computes the required physical quantities for performance evaluation. The obtained physical quantities are refined through a proposed computational method to extract the objective data necessary for multi-objective optimization. The second aspect is the multi-objective optimization of coil dimensions using Pymoo. Pymoo is a well-established open-source multi-objective optimization framework. It enables designers to define problems, establish constraints, and formulate quantitative equations, thereby integrating and binding optimization cases with real-world applications. By leveraging this framework, the desired optimization design is effectively achieved.