Abstract: To explore the impact of smart education workbench forms on users'affective cognition and develop a product form optimization method integrating affective evaluation, this study first employs Kansei Engineering theory. Using morphological analysis, the workbench form items and categories are deconstructed, and product samples are coded. Subsequently, the SelfAssessment Manikin (SAM) is utilized to measure Pleasure-Arousal-Dominance (PAD) three-dimensional affective data, with cluster analysis applied to identify affective cognitive groups. Finally, a Naive Bayes-based affective cognition prediction model for product forms is established. The exhaustive method selects optimal form layouts excelling across all three PAD dimensions from numerous solutions, and model accuracy is validated. Results demonstrate the rationality and reliability of this approach. The proposed PAD-Naive Bayes affective cognition prediction model can effectively forecast users'affective responses without relying on large-scale datasets, helping designers analyze the influence patterns of form elements on affective cognition and providing theoretical support for optimizing product form layouts.