Given that the current personalized content delivery methods exceed a duration of 3.5 h, this study delves into personalized content delivery for MOOC English education through the fusion of information from multiple sources. By synthesizing credible factors from various sources of information, a method is devised to distinguish the credibility of evidence. The Dempster combination rules are utilized to derive the final synthesis outcome based on the evidence sources. Utilizing the evidence sources from the synthesized resource information, the correlation of feature information is calculated by pairing each potential association, thereby establishing connections between all resource information. The resource delivery model aims to construct a subgraph of test questions using the learning behavior data collected from students during their educational activities. This entails the creation of a knowledge map of learning and amalgamating cognitive diagnostic data from diverse students to personalize content delivery. Notably, experimental findings demonstrate that the resource delivery time for the experimental group is under 3 h, marking it as the shortest duration among the three groups. In practical scenarios, selecting appropriate delivery strategies based on specific requirements and contexts is possible.