This study investigated multimodal multiple text comprehension (MMTC) among forty-eight Taiwanese university students, considering their cognitive and affective characteristics. A randomized trial with pre- and post-tests design was applied. The experiment group received a browser-embedded note-taking device along with epistemic prompting. The control group was provided with the browser-embedded note-taking device only. The results showed that situational interest, prior knowledge level, and cognitive load were associated with learners' MMTC. Besides, prior knowledge and cognitive load moderated the effect of epistemic prompting on MMTC. Students with low prior knowledge scored significantly lower in MMTC without graphics than those with medium or high prior knowledge in the control group. Moreover, the cognitive load had a stronger negative influence on participants' MMTC without graphics in the control group, but the negative impact was negligible on those in the experimental group. The contributions and implications of the findings were discussed regarding the MMTC intervention and instruction.