湿地面积不断萎缩导致自然湿地保护区生态功能持续减退，因此预测湿地面积变化对保护自然湿地具有十分重要的意义。本文以中国吉林省莫莫格湿地为研究对象，在分析影响湿地面积变化的主要因子基础上，考虑因子对湿地面积影响程度的差异，进行湿地面积预测。采用的预测方法为能较好地适应多因子预测问题的高度非线性特点的模糊推理方法，同时采用常用的多元回归分析方法进行莫莫格湿地面积预测，并将预测结果与模糊推理方法预测结果进行比较。研究结果表明：模糊推理方法模拟预测得到湿地面积合格率和预测精度均高于多元回归分析方法。该方法可在工程领域的多指标非线性预测、评价、聚类等研究中尝试使用。

The continuous shrinking of wetland area will lead to the continuous decline of the ecological functions of natural wetlands. In order to help wetlands to maintain a certain area, we should predict the natural area of wetlands first. In this paper, we used Momoge wetland in Jilin Province of China as the object of study to predict wetland area. First, we analyzed the main impact factors and their different influence on the area of Momoge wetland. We used the method of fuzzy reasoning method, which could well adapt to the highly nonlinear characteristics of multi-factor prediction. The results showed that the prediction accuracy of the fuzzy reasoning method is better than that of the multiple regression analysis method. The fuzzy reasoning method can be used in the studies of multi-index nonlinear prediction, evaluation, and classification in many engineering fields.

水利部公益性行业专项经费项目（201401014）；中央高校基本科研业务费专项资金项目(DUT14QY11)