在大坝变形监测中，需利用阈值法对数据信号进行小波去噪，而传统阈值法存在很多缺点。在小波去噪理论的基础上，结合Grubbs准则，提出了一种改进阈值门限的去噪方法，并用MATLAB编程实现了模拟信号以及大坝变形监测数据的去噪处理。实践表明，经小波改进阈值算法去噪后的数据均方根误差相对于改进前有所减低，信噪比也明显提高，改进效果明显。“,”Although the threshold method is usually applied in wavelet de-noising for data signal in dam deformation monito-ring, the traditional method has many deficiencies. In the combination of Grubbs rule, a new improved algorithm which changes the threshold is put forward on the basis of wavelet de-noising theory. As a result, the noise reduction processing for analog sig-nal and dam deformation monitoring data is realized by MATLAB programming. The practice shows that signals processed by the presented method have lower root mean square and better SNR, proving that the improvement effect is obvious.