The purpose of this study was to identify the relationship between human performance and heat map entropy in a system state judgment task through the visual interface screen. A prototype of an EID-based accident response support system was used as the visual interface screen for the experiment. Sisteen subjects performed an experiment in which the system state was judged based on the problems presented by the experimenter through the visual interface screen. The experimental results confirmed that the heat map entropy increased as the time to judge the system state increased. In addition, it was confirmed that the heat map entropy of cases that did not correctly judge the system state was significantly greater than that of cases that correctly judged the system state. It was found that a close relationship existed between the performance of judging the system state from the visual interface screen and heat map entropy. This means that it can predict human performance based on heat map entropy. Therefore heat map entropy can be used to evaluate the suitability of the visual interface design.