The collection and analysis of life data is the basis of reliability quantitative evaluation. This paper mainly discusses the determination of life distribution type and its parameter estimation. Because the life test is expensive and time-consuming, the test often has to wait until all the samples are unqualified. In addition, the observed values may be lost in the middle of field data, so the obtained life data are often incomplete samples. In mathematical statistics, there are special methods to deal with the parameter estimation and distribution test of this kind of incomplete samples, among which the result is the simplest when the life distribution is exponential (see statistical analysis of life data).