While many researchers in science education have argued that students’ epistemological understanding of models and of modelling processes would influence their cognitive processing on a modelling task, there has been little direct evidence for such an effect. Therefore, this study aimed to investigate the relation between students’ epistemological understanding of models and modelling and their cognitive processing (i.e., deep versus surface processing) on a modelling task. Twenty‐six students, working in dyads, were observed while working on a computer‐based modelling task in the domain of physics. Students’ epistemological understanding was assessed on four dimensions (i.e., nature of models, purposes of models, process of modelling, and evaluation of models). Students’ cognitive processes were assessed based on their verbal protocols, using a coding scheme to classify their types of reasoning. The outcomes confirmed the expected positive correlation between students’ level of epistemological understanding and their deep processing (r = 0.40, p = .04), and the negative correlation between level of epistemological understanding and surface processing (r = −0.51, p = .008). From these results, we emphasise the necessity of considering epistemological understanding in research as well as in educational practice.