<p>Experience products refer to goods whose value is perceived and evaluated by consumers through participatory experiential processes. Quantifying the perceived quality of experience products is of great significance for both content acquisition decisions on the platform side and product development process optimization on the enterprise side. Existing methods often struggle to address the ambiguity, correlation, and dynamics of perceived quality features for experience products. This study proposes a perceived quality quantification method based on user-generated content, comprehensively considering the ambiguity, correlation, and dynamics of features. First, useful online reviews are identified through data cleaning and usefulness screening. Second, driven by large language models, a model for perceived quality feature identification and network description is proposed. Finally, perceived quality quantification is conducted by extending the Ising model. The effectiveness of the proposed method is validated on three animation datasets, a micro-drama dataset, a new energy vehicle dataset and Yelp public dataset, and the commonalities and differences of its perceived quality features are analyzed and compared, providing directions for quality improvement. Additionally, its effectiveness and superiority are further verified by comparing it with the baseline method in two dimensions. Moreover, this method is also applicable to other functional and experience products.</p>