In metal fatigue tests, one or more outliers are observed at some time. It is hoped that the outliers are detected and can be rejected directly, but it’s difficult. Most papers about test data analysis just provide some brief criteria which are based on statistics. For better handling metal fatigue test data, this paper collects five rejectioncriteria and presents some derivation and formulas. And it’s useful to get a better understanding of these methods for test data analysts. At the end of this paper, a summary is given, some advantages and disadvantages of these methods were introduced, and suggestions are also provided.