Restricted by real condition, small sample method is usually used to measure the aircraft loads spectrum. The small sample data is a simple sample from the real matrix, which can-t ensure the adequacy and the integrity of data. The traditional average statistics method which is the common method to be used to develop loads spectrum can-t compensate for the defects of data. In order to overcome the shortage of small sample data, kernel density estimation was used to static loads spectrum, and better results were gotten. The basic theory of kernel density estimation was introduced and using the statistic of an airplane-s sinking velocity spectrum as an example to describe the mathematical course of kernel density estimation method on load spectrum statistics datailedly. The results show that kernel density estimation method has a good effect for restoring the original appearance of the load spectrum and can make up the shortage of small sample data.