This research investigates the use of natural stones as a sustainable construction material, leveraging their low embodied energy and local availability. Traditional stone masonry relies on skilled craftsmanship, which is increasingly scarce. To address this challenge, we develop a robotic planning system for autonomous stone layout, focusing on geometric optimization and structural performance. Our method translates traditional masonry principles—such as dense packing, joint interlocking, and course horizontality—into a multi-objective optimization framework for stone selection and placement. The algorithm minimizes gaps, maximizes stability, and ensures feasible configurations through discrete convolution techniques. Additionally, we employ a microscale rigid block modeling approach to assess structural behavior, simulating wall behaviors under shear-compression loading. In a case study, we virtually construct two 0.7 m × 0.7 m × 0.4 m walls using digitized stones, comparing layouts generated with full and partial optimization. Results show that walls built with the complete algorithm exhibit comparable force capacity and stiffness to mason-built walls, while partially optimized walls demonstrate reduced performance. This work advances robotic stone construction by enabling high-fidelity microstructure generation and load-bearing assessment, promoting sustainable building practices through automation.