Abstract: Outlier detection is one of the key processes to data preprocessing of satellite gravity gradiometry(SGG),and its objective is to obtain high quality SGG data.The performance of thresholding,Dixon test and wavelet algorithm was discussed in the outlier detection of SGG data.And two combined schemes were presented to overcome some deficiencies of these three methods,namely the combination of Dixon test and wavelet algorithm,and that of Dixon test,thresholding and wavelet algorithm.The simulation results show that both the two combined schemes can improve the quality of outlier detection effectively.