Abstract: It is necessary to choose an appropriate statistical method to analyze the results of the inter-laboratory comparisons. In this article, it is compared firstly that the process of calculation of the two sorts of common statistical methods: robust statistics and classical statistics, then the writers analyze the results of some inter-laboratory comparisons adopting this two sorts of statistical method respectively, using Grubbs test to judge the outlier, and the proper applying range of this two statistical methods is discussed: when there is much more data in comparison, it is suitable to analyze the results adopting robust statistics; at that time the data is limited, it is necessary to get rid of the outlier using Grubbs test method, paying attention to the maximum permitted error in the standards of the method or the product, after this, it is appropriate to analyze and appraise the remainder adopting classical statistics.