In recent years, many researches have explored the relationship between overt and covert motor activity (i.e., mental simulation). Consistent evidence has been provided in favour of close similarities between the two functions, particularly based on behavioural and neuroimaging studies on healthy participants. Interestingly, literature on the pathological population remains largely controversial. Yet, a clear understanding of whether and how mental simulation is modified by overt motor disorders is far from a speculative question, especially in view of the increasing interest for the use of mental practice in motor rehabilitation. Here, we explored whether a single set of cognitive skills is applied while solving tasks that implicitly require mental simulation of an action, or whether alternative strategies might be elicited according to the imager's actual motor capabilities. For this purpose, we recruited a group of patients who suffered from a stroke affecting selectively either the right or the left hemisphere, responsible for motor impairments ranging in severity. We required them, and a group of age-matched healthy controls, to perform a task of simulated grasping, and two tasks involving handedness judgments (on hands and gloves, respectively). Dissociations were found between the performances of patients suffering from left versus right brain damage, according to the task and, interestingly, the actual state of the imager's motor capabilities. This finding suggests that motor imagery might include alternative mental strategies that are independent from the actual state of the motor system. We discuss how these mental operations are differently affected by motor impairment, and consider the implications of the present theoretical finding for neurorehabilitation.