"The study of scenery preference as a psychological process has been a majorresearch topic in Environvmental Psychology. However, research on thistopic has emphasised the study of the scenery preferance as applied to natu-ral landscapes. This research ( Kaplan and Kaplan, 1982; Herzog andsmith, 1988 ) has proved that there is no universal relationship betweenphysical features and psychological processes. On the contrary, it seemsnecessary to explore the experience of the scenery as a whole and its assess-ment as a function of the subject's involvement in it. In spite of the theo-retical import of this line of research, there are few empirical studies withurban scenes.Our research study is intended to determine the preference criteria in theassessment of 16 urban scenes from the city of Madrid found in a sample ofresidents in this city. The 16 scenes were selected according to the two fol-lowing physical features: Presence / absence of "nature" (trees, lawns,hedges,. . . ) and open vs. closed appearance of the scene. Subjects' re-sponses were variables similar to those employed by Herzog and Smith(1988) "mystery", perceived physical danger, perceived social danger,shadows, nature, and depth.Results suggest the need to analyse the experience of the city's architecturalshapes as a whole as well as the difficulty of assessing a single physical orarchitectural feature in isolation. We therefore underline the importance ofassessing "urban scenes" in addition to natural landscapes. Furthermore,our results show that preference should be considered in the context of thepsyshological experience of feature. Finally, our study stresses the impor-tance of "mystery" and "shadows" as appropriate predictors of preference inurban scenes."