中文摘要：

本文以云南为例进行生态系统安全评价的研究.根据相关文献及云南实际,构建由自然、经济、社会三个子系统组成的生态系统安全评价系统,每个子系统包括6个指标,共有18个指标.运用熵权法和BP神经网络模型相结合的方法,对云南2012-2014年生态系统安全程度进行计算,结果表明：云南2012-2014年间生态系统安全程度不断提升,自然子系统安全程度由Ⅱ（不安全）上升为Ⅲ级（临界安全）,经济子系统和社会子系统均由Ⅲ级（临界安全）上升为Ⅳ级（较安全）.针对现状,提出云南生态系统安全程度提升的三点建议.

英文摘要：

This paper studies the ecosystem security assessment taking yunnan province as an example. Accord- ing to the related literature and actual situations in Yunnan Province, the authors built an assessment system, composed of natural, economic and social aspects, each of which includes 6 indexes, thus 18 indexes form the e- cosystem security assessment index system. The entropy method and BP neural network method are combined to evaluate the ecosystem security from 2012 to 2014 in Yunnan Province. The results show that ecosystem security in Yunnan Province has greatly improved： natural subsystem security has improved from Grade II （unsafety） to Grade III （criticality safety） ; economic and social subsystem security have improved from Grade III（ criticality safety） to Grade IV （good safety）. According to the current situation, three suggestions are proposed to improve the ecosystem security in Yunnan Province.