Visualisation has been used for decades as a strategy to help readers construct meaning from reading passages. Teachers across the globe have introduced visualisation mostly to primary students with native language background. They used the strategy to understand their own language. Little is known how this strategy works for university students who learn foreign language. Visualisation can be done internally (by creating mental imagery) and externally (by drawing visual representation). The product of visualising texts by using both models can be further investigated to find out if the meaning represented is appropriate to the meaning written in the text. This study therefore aims at exploring meaning by analysing the visual representations drawn by 26 English Education Department students of Syiah Kuala University after they read a narrative text. The exploration was conducted by looking at the image-word relations in the drawings. To do so, we consulted Chan and Unsworth (2011), Chan (2010) and Unsworth and Chan (2009) on the image-language interaction in multimodal text. The results of the analysis have found that the equivalence, additive and interdependent relations are mostly involved in their visual representations; and these relations really help in representing meanings. Meanwhile, the other three relations which are word-specific, picture specific and parallel are rarely used by the students. In addition, most students created the representations in a form of a design which is relevant to represent a narrative text. Further discussion of the relation between image-word relations, types of design and students' comprehension is also presented in this paper.