It has been argued that executive tests should capture central aspects of executive functions in everyday life such as initiating and monitoring parallel actions in low-structured environments (so-called multitasking; see Burgess, 2000). We present a cooking task in order to assess executive function impairments in brain-damaged patients, which focuses on a central feature of multitasking, the interleaving of tasks (Burgess, 2000). Behavioural performance of 21 brain-damaged patients (stroke, traumatic brain injury) and of a group of matched controls was analysed on the basis of a standardized protocol. In comparison to controls, the patients explored less, were less successful in monitoring their actions and corrected errors less efficiently. Interleaving of actions was observed less frequently in patients, with respect to both cooking itself as well as to subordinate goals (e.g., cleaning up). Interleaving proved efficient, as it was associated with less time to complete the task. Patients’ scores in the cooking task correlated with performance in both the Behavioural Assessment of the Dysexecutive Syndrome (BADS) Zoo Map Test and the BADS Six Elements Test, but not with tests of attention, verbal memory, or figural fluency, thus demonstrating convergent and discriminant validity. In summary, our task demonstrates that cooking can provide a valid testing ground for assessing a central aspect of everyday multitasking demands, namely, the interleaving of actions.