This paper presents a methodology for the optimization design of a fully parametric working attachment of a hydraulic excavator with the weight reduction and maximum stress minimization as the design objectives. A novel 3D force output model named limiting theoretical digging capability model is proposed to completely and accurately calculate the maximum force and moment an excavator can apply in the preferred bucket force space. On the basis of this model, the joint forces acting on the working attachment are calculated. Further, the critical digging condition and design loads used for the structure optimization can be determined by using the stress-based evaluation method. A fully parametric geometry model of the working attachment is constructed to achieve automatic design optimization and reduce the remodeling time. The plate thickness parameters and some critical structure parameters are selected as the design variables for the optimization design problem, which aims at minimizing the maximum stress in the structure and reducing the weight of the working attachment. Mixed variable genetic algorithm is used to solve this optimization problem due to its design variables include both discrete and continuous parameters. A simulation and optimization framework is established to support load calculation and processing, stress analysis, and structure optimization. Finally, the capabilities of the proposed methodology and established framework are demonstrated by a practical design. The results show that both the weight and maximum stress of the working attachment have an obvious decrease. This work provides a pragmatic tool for efficient virtual design and automatic optimization for the working attachment of an excavator.