杉木是我国南方最重要的速生丰产树种，在人工林中具有十分重要的地位，但是杉木种子涩籽率高是影响杉木产量和发展的主要原因之一，因此，积极探讨杉木种子涩籽的预测与防治是杉木种子生产中亟待解决的问题。而不同的地域是造成杉木种子涩籽量差异的重要因子之一，为了进一步探讨杉木种子涩籽在地理上的流行规律，本文试图运用一种新的方法———人工神经网络方法对杉木种子涩籽与地理之间的关系进行研究，建立了杉木种子涩籽地理流行ＢＰ网络模型。结果表明：所建立的ＢＰ 模型对模拟预测不同地域杉木种子涩籽的涩籽率具有较高的精度，平均模拟精度为８８－４０ ％ 。这不仅为杉木种子园的合理布局提供理论依据，而且也为人工神经网络在林业科学研究中的应用开辟新的思路 Cunninghamia lanceolata is one of the most important fast-growing and high-yielding species in southern China, and plays a very important role in plantations. However, the high rate of astringent seeds in Chinese fir is one of the main factors that affect the yield and development of Chinese fir. Therefore, And prevention and control are the problems to be solved urgently in the production of Chinese fir seed. In order to further explore the geographically prevailing laws of Cunninghamia lanceolata astringent seeds, this paper attempts to apply a new method --- Artificial Neural Network Astringent seed and geographical relationship between the research to establish a Chinese fir seed Sekkisei geographical popular BP network model. The results showed that the established BP model had higher accuracy for simulating the astringent seed rate of Chinese fir seed astringent seed, with the average simulation accuracy of 88-40%. This not only provides theoretical basis for the reasonable layout of Chinese fir seed orchard, but also opens up new ideas for the application of artificial neural network in forestry scientific research