A multi-objective optimization method considering duty cycle suitability is developed for collaborative robot joint motors, addressing acceleration demands and repetitive duty cycles. The motor’s operating envelope is discretized into a structured grid using trajectory-based statistics. Grid cells are statistically weighted, and representative points are selected to capture typical operating conditions; sparsely populated regions are consolidated to reduce finite-element simulation demands. This enables efficient optimization to minimize weighted loss and rotor inertia under practical constraints, solved via the Non-dominated Sorting Genetic Algorithm II. Thermal safety is then verified through transient temperature-rise simulation, and the final designs are selected based on torque-ripple performance. Compared with the original motor design, the optimized motor achieves over a 50% reduction in weighted loss, ~ 25% lower rotor inertia, and ~ 40% reduction in torque ripple, accompanied by higher efficiency and reduced temperature rise. Additionally, compared to a motor optimized solely for rated-point efficiency, the proposed method results in lower losses and heat generation over actual operating cycles, demonstrating the practical effectiveness of this optimization approach.