Smart product design needs to take into account both functionality and aesthetic needs, and traditional design methods are difficult to balance these multidimensional goals. This study integrates cognitive science and machine vision technology to construct an intelligent product design optimization method. In the method, user needs and aesthetic forms are regarded as complex adaptive systems, Gray coding is used to encode product features, a multi-objective optimization model is constructed based on NSGA-II algorithm, and the optimal design scheme is selected through non-dominated sorting and congestion calculation. The experimental design invites 30 designers to participate and provides three types of external incentives, namely, far-domain incentives, near-domain incentives and constraints. The results show that the far-field incentive group generates an average of 3.61±1.73 number of solutions in the design stimulus phase, which is significantly higher than the 2.5±1.52 of the near-field incentive group and the 2.44±1.47 of the constraints group. In the evaluation of the importance of technological attributes, the TA2 proximity coefficient reaches 0.6473, which is the second highest, while the proximity coefficients of TA8 and TA14 are both 0.0001. It shows that the impact of different technical attributes on user satisfaction varies significantly. The study shows that the design optimization method integrating cognitive science and machine vision can effectively improve the usability of intelligent products, resolve technical conflicts, and realize the innovative design of products oriented to user needs.
