Our previous work demonstrated that the action encoding parietofrontal network, which is crucial in planning and executing motor tasks, is less active in prosthesis users who imitate movements of intact actors (mismatched limb) versus prosthesis users (matched limb). Such activation could have behavioral consequences in prosthesis users rehabilitating with intact therapists. The goal was to identify behavioral effects of matched versus mismatched limb action imitation in naïve users of prostheses. Intact subjects donned a specially adapted prosthetic device to simulate the wrist and forearm movement that transradial amputees experience. While electrogoniometry was recorded, non-amputated prosthesis users (NAPUs) observed and imitated demonstrations of a skillful motor task performed by either an intact actor or NAPU. We hypothesized that NAPUs would elicit less motion variability when performing matched versus mismatched imitation. Matched imitation resulted in a significant decrease in shoulder motion variability compared with mismatched imitation. The matched group also developed elbow motion patterns similar to the NAPU demonstrator, while the mismatched group attempted patterns similar to the intact demonstrator. This suggests a behavioral advantage to matched imitation when adapting to a prosthetic device, as it yielded more consistent movements and facilitated development of new motor patterns. Further, these results suggest that when prosthesis users are faced with the impossible task of imitating movements of an intact hand, they perform this action with greater variability and poorer technique. This work has implications on how prosthetic device operation is conveyed to persons with amputation as their clinical interactions often involve mismatched limb imitation.