Instructional video has become an important teaching resource. In related studies, researchers have found that the playback speed of instructional videos affects learning effect. However, it is not clear whether the learning effect obtained when learners choose different playback speeds is affected by the interactive design of the video (e.g., such as pre-embedded questions). Therefore, in this study, behavioral tests and eye-movement techniques were used to investigate the learning outcomes and cognitive processing of 90 participants based on instructional video playback speed (1.0×, 1.25×, and 1.5×) and question embedding types (no-question embedding and pre-embedded questions). The results showed that the pre-embedded questions not only improved learners' attention to the main content of the video, but also enhanced learning performance. An appropriate increase in playback speed (1.25× speed) is beneficial to learners' cognitive processing and learning depth, but excessive speed (1.5× speed) can be detrimental. Pre-embedding questions can broaden the range of speed acceptable to learners while maintaining learning efficiency. There was no significant difference in learning outcomes between 1.0, 1.25, and 1.5 times speed when questions were pre-embedded. However, the learning effect of 1.5 times speed was considerably lower than that of 1.0 times and 1.25 times speed when no questions were embedded. This study not only provides a new research perspective for the related research of instructional videos, but also provides a reference for the design and development of teaching video, so as to meet learners' individual needs and improve the effectiveness of online learning.