The dual-task paradigm represents an established laboratory approach to investigate multitasking. Hereby, performance differences as a result of executing (at least) two tasks simultaneously relative to each single task are considered to indicate multitasking abilities. However, this method has been criticized for its weak ecological validity as it hardly operationalizes everyday multitasking. A solution for this may be to combine several subtasks such that together they represent a more naturalistic multitasking scenario. The Multi-Attribute Task Battery (MATB) simulates such a scenario requiring individuals to steer a joystick, to respond to radio calls and emergency lights, and to manage the fuel tank of an aircraft. But it has not yet been established whether a multitasking scenario, such as the MATB, indeed predicts naturalistic multitasking performance better than the dual-task paradigm approach. We addressed this issue by investigating whether naturalistic multitasking performance of 25 volunteers from the Universität der Bundeswehr München was better predicted by MATB performance than laboratory multitasking. Laboratory multitasking was operationalized by means of performance decrements in solving math equations and responding to radio signals in dual- relative to single-task conditions. Naturalistic multitasking was operationalized by means of decrements in shooting performance in triple- and dual- relative to single-task conditions. Bayesian regression analyses indicated at least moderate evidence in favour of correlations between MATB performance and laboratory and naturalistic multitasking, but not between the latter. Thus, our data suggest that a multitasking scenario, such as the MATB, might be more applicable to predict naturalistic multitasking performance than dual/multi-task decrements.