The attitude monitoring of the wind turbine tower is an important part of the safety monitoring of the wind turbine tower. In order to solve the problem that the data noise is generated in the process of attitude monitoring and it is difficult to determine the operation attitude of the tower, the PCA algorithm based on SVD and the EMD algorithm are integrated to purify the noisy data. The noise generated in the process of attitude monitoring is composed of periodic component and random high frequency noise. Firstly, SVD-PCA algorithm is used to extract the main intrinsic frequency of the data, and the intrinsic frequency of the periodic interference data is reconstructed according to the characteristic frequency. Then EMD algorithm is used to decompose and reconstruct the data to eliminate the periodic characteristics of the data. The validity of the algorithm is verified by simulation signals. Taking a wind power tower in Yancheng as the research object, the algorithm is applied to the filtering of long-term monitoring data of tower top inclination sensor, and good results are achieved. After judgment, the wind tower is still in safe service condition.