This study explores the construction of traditional culture English learning resources through the integration of data mining technology, Bayesian neural networks (BNNs), and Dragonfly optimization techniques. The aim is to develop culturally authentic and linguistically rich educational materials that enhance language proficiency while deepening learners' understanding of traditional cultural nuances embedded within the English language. By leveraging BNNs, which account for uncertainty and variability in traditional cultural texts, and employing Dragonfly optimization to optimize search algorithms, educators can curate personalized and adaptive learning resources tailored to individual learners' needs and preferences. The utilization of data mining technology facilitated the extraction and analysis of vast datasets encompassing diverse aspects of traditional culture, including texts, images, and multimedia content The innovative approach presented in this study promises to revolutionize language education by fostering a deeper appreciation for traditional culture and promoting cross-cultural understanding and communication.