In this study, we developed an intelligent food packaging design and evaluation framework integrating AI-generated design, sensing materials, craft aesthetics, and a multi-attribute decision model. A craft-art dataset containing patterns, colors, and compositions was first established. Visual features were extracted using a Convolutional Neural Network, whereas Generative Adversarial Network and Latent Diffusion Model algorithms generated diverse patterns that preserved cultural symbols while extending modern visual appeal. Three sensing materials, namely, temperature–humidity sensing films, photochromic films, and eco-friendly substrates, were experimentally tested for response time, color difference (ΔE), mechanical strength, and food safety to ensure real-time responsiveness and stability. A user study with 30 participants compared three prototypes: (1) conventional packaging, that is, packaging without AI-generated design or sensing-interactive functionality, (2) algorithm-generated packaging, and (3) algorithm sensor-interactive packaging. Participants evaluated visual aesthetics, interactivity, and cultural identity. By using Analytic Hierarchy Process and Fuzzy-Technique for Order Preference by Similarity to Ideal Solution for multi-attribute analysis, the sensor-interactive packaging achieved the best performance across Structural Similarity Index Measure, Frechet Inception Distance, color harmony, and user rating (correlation coefficient = 0.670), outperforming the others. The findings confirm that integrating AI-driven generative design with sensor-based materials enhances the aesthetic and interactive experience of packaging. The proposed framework provides a quantifiable and practical evaluation model for cultural-creative design, offering broad applicability to food, cultural, and sustainable packaging industries and supporting the advancement of smart packaging and cultural innovation.