Visual context, such as an association between a target location and a distractor configuration, has been shown to be implicitly learned through repeated experiences in a visual search task, and facilitates the search performance (contextual cueing effect). In the present study, we examined whether the whole layout or only the local half-layout contributed to contextual cueing. In a training session, the whole layout or the local half-layout including the target location was repeated. Then in a test session, half of the layout including the target location was presented while the remaining half of the layout was randomly changed. The results showed that the search performance in the repeated whole layout condition was equivalent to the condition where only half of the layout was repeated throughout both the training and the test sessions. This suggests that the local layout around the target location was selectively learned and used for contextual cueing even if the whole layout was repeated.