Based on education agent application, the paper designed emotional detection model and overall schema of detection
for learners. Then the paper completed personality learning monitoring for learners and defined emotional attitude,
cognitive state and learning preference for learners through extraction and analysis on emotional information between
education agent and learners. The paper conducted personalized learning resources service for learners by using automatically
generated courses tool based on emotional detection results in order to improve human-computer interaction status
in remote learning and improve overall learning effect. The paper also studied on evaluation method of human-computer
interaction system and focused on the multi-objective decision-making method. Then the paper realized more accurate
quantitative evaluation through evaluation of NAP on human-computer interaction.