Abstract: As an important statistic, the residual error of adjustment is the estimation of the true error of observation in nature, but quantitatively it is a linear function of a group of true errors, whose values reflect the corporate effects of many true errors.By considering the quantitative relation between the residual error and the true error, this paper focuses on the data snooping and the iterative methods of weight defining for the robust estimation which are used to detect gross error.We point out the problems that will be met during the course of substituting the true error for residual error and how to resolve them.The statistical feature of mean square of weight is also discussed in this paper, so we deduce the estimation of the variance of weight and the statistical distribution of the estimation of mean square of weight, and also prompt the mathematical features of the latter.The statistical analysis of adjustment precise system is discussed.