In the case of big data sample, data reliability could be evaluated easily by probability and some other methods. In engineering practice, the credibility assessment of small sample data without reference standard is often encountered, which is a hard problem. A method of credibility evaluation for small sample data based on grey system theory and cloud model is proposed. In accordance with the grey approaching correlation, the grey estimated value of small sample data is calculated. On the basis of grey estimated value, the parameters of mean, entropy and hyper entropy for cloud model can be calculated. Then the cloud model is built. Take advantage of the cloud model, amount of cloud drop data is produced around the data of the small sample. In the light of the confidence degree of cloud data, the credibility of each data for small sample data can be obtained. Finally, several simulations and analysis are conducted. The simulation results demonstrate the feasibility and effectiveness of the proposed method.