In the absence of effective drugs to control the COVID-19 epidemic, the main intervention of human factors, namely strict isolation, may be the best prevention and control method at present. The conclusion of data empirical study using Benford's Law is of great significance.Research purpose of this paper analyze COVID -19 data to predict the authenticity and reliability, and on this basis, the method is to use Benford's Law and the panel model for 51 countries or regions COVID -19 data statistical analysis, the results of the study found that "other areas" unreliable data, Australia, Pakistan and global data are greatly influenced by artificial factors, Africa, Oceania data with several other states Data according to have significant difference, compared the southern hemisphere and northern hemisphere, the first phase of the data and the second stage also has the obvious difference between the data, The COVID-19 data are also predicted to suggest that the outbreak may have multiple iterations.In conclusion, in most cases, when COVID-19 data deviates from Benford's Law, epidemic prevention and control is better; otherwise, it is worse.