In order to solve the problems in the estimation of industrial enterprises under the designated size, the existing weight adjustment range is expanded to improve the estimation accuracy of industrial enterprises under the designated size. On the one hand, it solves the problem of unnatural extinction of catalog enterprises. The unnatural extinction of sample units and the unnatural extinction of sample layer are discussed respectively. The unnatural extinction is regarded as a unit without answer. The sample matching method is introduced to select the most ``similar" normal reporting enterprises to match with the unnatural extinction enterprises, and the weight of the unnatural extinction sample enterprises is adjusted to the normal reporting sample enterprises. On the other hand, the estimation bias of non-catalog enterprises is solved. The weight adjustment ideas based on superpopulation model estimation and inverse weighted estimation of propensity score are discussed, respectively. Linear and nonlinear models are selected for superpopulation model estimation. In the inverse weighted estimation of propensity score, the solution of propensity score is mainly studied. Based on generalized boosted model (GBM) algorithm, weight is introduced in the iterative solution process, and w-GBM algorithm is proposed. At the same time, a combined estimation method is proposed by weighting the logistic regression estimation in the parameter estimation method and the w-GBM algorithm or GBM algorithm in the nonparametric estimation method. The numerical results show that the ideas proposed in this paper are feasible.