To solve the problem that the factors, which affect the coalition efficiency in multi-agent systems(MAS), have
strong ambiguity and uncertainty, a novel method for the evaluation of agent coalition is proposed based on the cloud
model which is introduced as a transforming model between qualitative and quantitative. Contrast to the existing works,
the proposed method is more simple and effective, and easy to work in projects, which can deal with the ambiguity and
uncertainty in evaluation indicators, thus can provide more scientific and reasonable decision for choosing the best coalition
in some complex control systems.