Voice-driven adaptive cooperative systems have been widely used. However, the ambiguity of the voicebased task goals remains a challenge. How to integrate fuzzy task goals into adaptive cooperative systems to make decisions is a problem that should be resolved imperatively. We propose a fuzzy voice-driven selfadaptive cooperative framework and a corresponding decision method. Firstly, the system publishes tasks, submits tenders and makes decisions based on the marketbased approach. Secondly, multi-attributes evaluation information is transformed into 2tuple linguistic and the VIKOR method is used to rank the candidates. Finally, we take the heterogeneous robot self-adaptive cooperative system as an example to illustrate the decision-making steps of the proposed method, and the experimental results verify the effectiveness of the method.