Abstract: In this paper, we used the time series method to forecasting population fluctuation of Bandicota indica，the main pest rodent in rice fields of Pearl River Delta，Guangdong province．Based on the data surveied in Liboshui district of Shiwan town，Boluo county，Guangdong province，from November of 1991 to December of 1993；we built a model by using triple exponential smoothing model combined with seasonal adjustment modeI，the smoothing constant was fixed as a=0.02, to forecast the population density month by month of 1994 continuously，the rate of accuracy of forecasting was 81.93﹪ ．Based on the data from November of 1991 to December of 1994．we forecasted the population density of 12 months of 1995 continuously，the rate of forecasting accuracy of two years was 86.56%.

Key words: Bandicota indica, Time seris, Forecasting

摘要： 本文利用时间序列方法(三次指数平滑法)，并结合季节指数法，建立了板齿鼠种群数量中长期预测(6个月～1年)的时间序列模型，依据1991年11月~1993年l2月在广东省博罗县石湾镇里波水管理区对板齿鼠种群数量动态的调查数据．分别预测了1994年和1995年板齿鼠种群数量的发生，预测准确率达到了86．56％。

关键词: 板齿鼠, 时间序列, 预测预报