Abstract: Content characteristics of silica dioxide in iron ores imported from 23 countries, totally 490 batches, in recent two years are investigated by statistical methods. Based on the study on data distribution characteristics, a new robust statistics method,kernel density estimation, coupled with bootstrap method is introduced to acquire the representative values of silica dioxide content and the standard deviations for imported iron ores. It is clearly demonstrated that this method shows a prior advantage to give a robust description in explanation for central tendency and variation of data profiles. It has also proved that the conclusion of substitution of bootstrap mean value and its standard deviation for the representative value of silica dioxide content from single specimen analysis is reasonable and effective.