Although objects with curved contours are generally preferred over those with sharp-angled contours, the strength of this preference varies according to several factors. In the present study, non-Western Japanese observers viewed and rated their preferences (e.g., liking or attractiveness) for real and meaningless objects with curved or sharp-angled contours. We varied the presentation time (90 ms vs. until a response was received) and the response measure (like/dislike vs. 1–100 rating scale). When using like/dislike ratings, a preference for curved objects was found only when images of real objects were presented briefly (90 ms), whereas this effect was reversed (i.e., increased preference for sharp-angled contours) when using the 1 to 100 scale under the until-response condition. In addition, the curvature effect was not observed for real objects when the like/dislike rating and the until-response condition were employed or when the 1 to 100 scale and 90 ms presentation time were used. The curvature effect for meaningless objects remained unstable regardless of presentation time or response measure. Similar to the preference for real objects, a preference for sharp-angled objects was observed when preference was measured using a 1 to 100 rating scale. Taken together, the present findings indicate that the preferences for curved objects were situation-dependent in Japanese observers.