Investigators have previously examined the effects of visuospatial working memory load on visual search using dual-task paradigm in which participants performed a visual search task concurrently with either a spatial working memory (SWM) task or an object working memory (OWM) task. It has been found that SWM load impairs the expression of the contextual cueing, not the learning of contextual cuing. In addition, the OWM load does not affect the learning and expression of contextual cueing. Meanwhile, some experiments have revealed that only SWM load affects the search efficiency in traditional visual search and real-world scenes search. The present article extended traditional visual search to real-world scenes search and investigated how the SWM load affected the learning and expression of contextual cueing in real-world scenes by employing dual-task paradigm. In the present study, the participants performed a single task or dual tasks. In the single task, participants were asked to search a specific target (“┫” or “┣”) from a series of displays, in which 50% displays were repeated and the other 50% displays were updated constantly in order to elicit the contextual cueing. In the dual tasks, participants were asked to perform a search task during the delay interval of spatial working memory task (remembering four spatial locations which were presented one by one). In Experiment 2a, sixteen participants performed a dual task and single task in the learning phase and the test phase separately. Conversely, in Experiment 2b, sixteen participants performed a single task and dual tasks in the learning phase and the test phase separately. However, in Experiment1, sixteen participants only performed the single task in both learning phase and test phase. Finally, we analyzed the results of each experiment and compared the standardized context cueing score (CC) of Experiments 2a and 2b with that of Experiment 1 respectively. The main findings were as follows: (1) In Experiment 2a, search advantage for repeated displays was observed in the presence of concurrent spatial working memory load. This advantage was observed in both the learning and test phases. And compared to the non-loads condition in Experiment 1, the CC were higher in both the learning and test phases; (2) In Experiment 2b, search advantage for repeated displays was only observed in the learning phase. The advantage disappeared in the presence of concurrent spatial working memory load in test phase. And compared to the non-loads condition in Experiment 1, the CC were only significant in the test, specifically the CC was decreased, but no difference in the learning phases. These findings suggest that the effects of spatial working memory load on the contextual cueing in search of real-world scenes are not same as those in the traditional visual search. Spatial working memory load would improve the learning, but impair the expression of the contextual cueing in search of real-world scenes. In other words, the spatial working memory load affects the learning and the expression of contextual cuing simultaneously. Meanwhile, the current study also suggests that spatial working memory load do not affect the nature of explicit learning of contextual cueing.