This study investigates methods and effectiveness of implementing Differentiated Instruction (DI) in intelligent education based on learners' learning needs. The paper employs a mixed-method approach to collect and analyze data from various learner groups. Guided by the assessment results of the KANO model, the research prioritizes attributes for intelligent enhancement during the learning process and devises appropriate methods for instructional interventions. Experimental results demonstrate that DI tailored to learning needs effectively enhances language teaching achievements. Specifically, in language education, regarding Performance Attributes (PA), meticulous segmentation of instructional content is crucial to correspond with specific language knowledge. For Threshold Attributes (TA), emphasis should be placed on fostering autonomous learning and adaptive communicative functions during the teaching process. Regarding Excitement Attribute (EA), it is essential to consider the distinctive needs of learner groups. Finally, the paper discusses collaborative issues among teaching teams, learners, and technical teams.