The present study aims to explore the impact of mismatch between learners' media preference and media assigned to them on their learning performance. The participants were 214 adult Chinese e-learners enrolled for an English course, and were assigned to 6 groups based on their media preferences and the media assigned to them. A 2 × 3 "ANOVA" analysis with dependent variable of learning results and independent variables of media preferences (auditory vs. visual) and media types (audio-only vs. on-screen text-only vs. audio and on-screen text combined) was performed for the results of the experiment. Results indicate that the mismatch between media preferences and received media did not bring the expected worst learning outcome. Moreover, the learners' media preferences did not lead to significant differences on the learners if they were provided with the same type of learning media; however, if different learning media were assigned to the learners with the same media preference, significant differences in learning outcome can be found. Another finding is that the combined medium has the most positive impact on the participants' learning performance. The questionnaire survey and the semi-structured interview conducted at the end of the experiment also revealed the popularity of the combined medium.