In this paper, we propose a time-series comfort evaluation system for streetscape by using electroencephalogram (EEG). Since streetscapes have an impact on people's psychological condition on a daily basis, a streetscape evaluation system is required for streetscape improvement. In conventional ways, questionnaires have been mainly used to evaluate characteristics of the streetscapes. However, they have two pernicious defects. The first is that people usually forget their feelings before answering questionnaires. The second is that it is difficult to detect the condition of environmental elements because streetscapes change constantly. To detect environmental elements which affect people, we tried to evaluate comfort in streetscape using EEG in addition to questionnaires. In the experiments, we measured EEG signals when subjects walk and view streetscapes of Japanese city. We classified EEG signals into two classes: “comfort” and “discomfort”. Through the experiments across twenty one subjects, classification accuracy was achieved 76.1% using non-linear SVM. The results showed that a time-series EEG system, outputs discrete value of comfort in viewing streetscape, was constructed. We conducted additional experiment to verify the effectiveness of the system. For the experiments, we could detect comfort and discomfort elements that subjects really feel. We confirmed that the method using questionnaires and EEG is effective to evaluate comfort that people feel in streetscapes.