The work aims to quantitatively study the key factors of artificial intelligence (AI)-generated cosmetic packaging design from the perspective of user perception evaluation to analyze the relationship between them and user satisfaction, and provide AI-based packaging design-related development strategies. Through literature research and user interviews, an evaluation system containing 5 dimensions and 18 factor indicators was constructed in combination with experts' opinions; A respondent questionnaire survey was conducted on AI-generated packaging designs based on the evaluation indicators; And an IPA model was used to analyze the degree of importance and satisfaction of each factor indicator of AI-generated cosmetic packaging designs. The research found that users showed high recognition and satisfaction with AI-generated packaging design in terms of information communication and creative appeal dimensions, while they generally tended to be dissatisfied with the functional usability and user experience aspects of the packaging. Therefore, enhancing the utility and user experience of AI-generated packaging is a key to improve user satisfaction. The effectiveness and user satisfaction of AI design can be enhanced through strategies such as training AI models with existing mature packaging public models, enriching data sets by enhancing multi-modal learning, optimizing the interpretability and controllability of the AI design generation process, and building a multi-stakeholder participatory platform for generative design.