Abstract: It is an important basis for evaluating the rational level of urban development to determine reasonably scale of urban land use.This paper chooses the important influence factors of developed areas：Annual GDP，the gross output of industry and agriculture，total population etc.，and makes use of grey GM（1，N），BP Neural Network to construct forecasting models，respectively，as well as analyzes and makes comparison of the advantages and the disadvantages of each model.This paper proposes the combination forecasting model by using standard variance to allocate the weight.The result of practical example shows that the forecasting accuracy of combining forecasting model is better than that of the single model，which is a beneficial exploration for the forecasting method of developed areas.

摘要： 合理确定城市用地规模，是衡量城市理性发展的重要依据。选择了年度GDP、工农业总产值、总人口等重要的建城区影响因子，分别用灰色GM（1，N）、BP神经网络构建了单项预测模型，并对各模型的优缺点进行比较分析。其次采用标准差法进行权重分配，建立组合模型。实例计算结果表明，组合预测模型的精度优于其他两个单一预测模型，这是对建城区面积预测方法的有益探索。