Elaborated Intrusion Theory (EI) postulates that imagery is central to craving, therefore a visually based task should decrease craving and craving imagery. This study provides the first laboratory test of this hypothesis in naturally occurring, rather than artificially induced, cravings. Participants reported if they were experiencing a craving and rated the strength, vividness and intrusiveness of their craving. They then either played 'Tetris' or they waited for a computer program to load (they were told it would load, but it was designed not to). Before task completion, craving scores between conditions did not differ; after, however, participants who had played 'Tetris' had significantly lower craving and less vivid craving imagery. The findings support EI theory, showing that a visuospatial working memory load reduces naturally occurring cravings, and that Tetris might be a useful task for tackling cravings outside the laboratory. Methodologically, the findings show that craving can be studied in the laboratory without using craving induction procedures.