Study of the student understanding of graphs in physics and mathematics was conducted on mostly fourth year students of the University of Science (45 students, preservice physics teachers) and Faculty of Humanities and Social Sciences and Centre for Croatian Studies in Zagreb (47 students of psychology). Constructed test consisted of four pairs of parallel questions which included exactly the same graphs: the context in one half of the questions was from kinematics (physics) and the other half of the questions was related to money (mathematics). The questions examined qualitative and quantitative understanding of the slope and area under the graph. The participants first solved the questions on the computer while their eye movements were recorded. Afterwards the same questions were solved on paper and along with the answer they had to provide explanations. Categorization of the explanations of the participants gave an insight in the most frequently used right and wrong strategies in solving the questions from the conceptual fields of graph slope and area under the graph. The results have shown that for students the idea of slope (derivation) is intuitively more understandable than the idea of area under the graph (integral) of the function, and the qualitative questions were easier than quantitative. Physics students were better than the psychology students in solving the questions about graphs in both contexts, however one part of them did not apply their developed procedures of solving questions about graphs on the analogue questions in new contexts, i.e. they solved the physics questions better than mathematical. On the contrary, students of psychology were better in solving qualitative assignments in the context of mathematics, regarding the everyday life, than physics, probably due to the everyday life experience. Question solving time was obtained with the assistance of the eye tracking device. Longer time of question solving suggested its higher difficulty because poorly solved questions took longer time, e.g. area questions took longer period than slope questions. The students were faster in solving the qualitative questions in the context of physics while quantitative questions took equal amount of time in both contexts. Further analysis of the question viewing and solving has shown that the physics students have analysed every single detail of the questions, while the psychology students only after a short analysis decided which answer was the right one. Furthermore, the students in general have taken longer time to view/analyse specific parts of the questions in the context of mathematics than physics which implies that it is easier to get the solutions when they can rely on solid formulas that they remember from the previous school period, as opposed to when they need to apply the knowledge in the specific situations and new contexts. Similar results were also obtained from the analysis of the students strategies. The results of this study imply that the transfer of knowledge between mathematics and science areas exist, but is insufficient, thus classes should be prepared, planned and conducted in order to increase the transfer. To reach that purpose we should insist on conceptual understanding and integration of content of each subject separately, yet also between subjects. We should insist less on the conclusions based on the formulas and encourage word elaboration for it develops critical and logical thinking, and the method of approaching complex problems, i.e. interactive enquiry-based teaching methods should be applied.