The formation of the digital divide is influenced by both objective factors, such as insufficient digital resources, and subjective factors, such as technology acceptance. This study employs a mixed-methods approach, utilizing the KANO model to analyze learners' demand attributes and the UTAUT model to examine subjective factors influencing technology acceptance. Standardized tests and survey questionnaires are used to assess digital learning outcomes. By collecting and analyzing data from diverse learner groups, this study aims to explore strategies for bridging the digital divide when transforming traditional online education into intelligent education, particularly in technologically and infrastructurally underdeveloped regions, including developing countries. The findings indicate significant differences in demand priorities among learners, as well as notable variations in how different learner groups classify their learning needs. Based on the demand analysis, targeted functional development can be implemented to reduce development costs in developing countries while maintaining the accessibility of digital resources. Moreover, user preferences for digital learning vary across groups; AI-driven identification and personalized recommendations can facilitate a more inclusive and equitable digital learning environment. Additionally, lowering the barriers to technology use through AI, enhancing engagement, and improving perceived effectiveness can significantly strengthen learners' confidence and motivation in bridging the digital divide. Finally, governments, educational institutions, and corporations should establish stronger communication and collaboration mechanisms to jointly address the pervasive digital divide in the era of intelligent education, particularly in developing countries.