变形监测一定程度上可以预测沉降发展趋势,本文采用灰色模型和神经网络的建模理论,介绍了GPS测沉降数据的预报处理流程,并利用上海市CORS网的数据分析GPS沉降预报,运用模型对GPS测沉降数据进行预测,最后利用中误差理论对精度进行评定,得出了灰色模型和神经网络模型在GPS沉降数据预报方面精度是可靠、精确的结论,从本文数据看神经网络预测的精度比灰色模型预测精度更符合实际。 Deformation monitoring can predict the trend of subsidence to some extent. In this paper, we use the gray model and neural network modeling theory to introduce the forecasting process of GPS subsidence data, and use the data of Shanghai CORS network to analyze the GPS subsidence forecast, GPS settlement data to predict, and finally use the error theory to assess the accuracy of the gray model and neural network model obtained in the prediction of GPS settlement data accuracy is reliable and accurate conclusion, from the data in this paper neural network prediction accuracy than Gray model prediction accuracy is more realistic.