Traditional methods such as scoring method, Likert scale method, etc, suffer from uncertainness, randomness and inaccuracy. To solve these problems, a cloud model is presented in this paper. Based on fuzzy mathematics and statistics, the cloud model can convert the qualitative evaluation of the product to quantitative scores. It’s a better way to overcome the disadvantages of direct scoring method that cause more inaccuracies. A prouct satisfaction evaluation model with three dimensions (i.e., evaluation index, product sets and estimator) is then constructed, which combines the cloud model and the grey relational analysis. Finally, the effectivity of the presented model is demonstrated by a case study.