<p>Chemistry learning and understanding depends on the ability to think circularly at three levels: the macroscopic, the sub-microscopic, and the symbolic level. This ability is mostly given for granted in higher education, although many students struggle in representing the sub-microscopic level, which is outside their range of experience. In addition, handling these three levels at a time may easily overload the Working Memory (WM), thus hindering meaningful learning. It has been demonstrated that success in chemistry learning correlates with good visuospatial abilities as well as with WM capacity. Nevertheless, the knowledge about specific cognitive processes involved in chemistry learning is still limited, especially concerning undergraduates. The present work contributes to this goal by investigating the differential involvement of verbal and visuospatial WM components in chemistry learning. We evaluated a sample of 315 undergraduates with a chemistry task (CT), and a subset (<i>N</i> = 81) with three different cognitive tasks (the Listening Span Test, the Imaginative Puzzle, and the Cattell’s Fluid Intelligence Test), aimed at assessing the verbal WM, the visuospatial WM and the estimated IQ, respectively. We sought correlations between WM components and chemistry performance: our results highlight a positive correlation between accuracy in performing the chemistry task and performance in the visuospatial WM task. Supplemental analysis compared the role of WM components in relation to the types of chemistry task questions (textual or visual). Our results showed that good visuospatial WM performance promoted the accuracy of students’ performance when they were asked to manipulate both textual and visual information; this advantage was enhanced when visual information was involved.</p>