我国图书馆普遍存在纸质图书质量欠佳与藏书利用率偏低的问题,理论与实践研究中鲜见卓有成效的解决方法。本文提出了预订新书的读者需求量的预测方法,以嘉应学院图书馆为例,通过线性回归分析、t检验和相关度分析三种方法的统计分析表明,预测结果与实际结果吻合良好。在预测读者需求量的基础上采用新的复本决策,结果表明,该决策方法能够大幅度提高藏书利用率和购书经费的使用效率。 The problems of the poor quality of paper books and the low utilization rate of books in our country are common in libraries, and few effective solutions are found in the study of theory and practice. In this paper, we propose a method to predict the readership of new books. Taking Jiaying University Library as an example, the statistical analysis of the three methods of linear regression analysis, t-test and correlation analysis shows that the prediction results are in good agreement with the actual results. Based on the forecasting of readers’ needs, the new replica decision-making method is used. The results show that the decision-making method can greatly improve the efficiency of collecting books and using books.