Abstract：To improve the predictive accuracy of Beidou satellite clock bias (SCB), it is necessary to adopt a proper method of data quality control to preprocess SCB. The phenomena consisting of phase jumps, outliers, and missing data exist in long term Beidou SCB sequence. Considering all these factors, a data quality control method for the Beidou-2 SCB series is proposed. The method is as follows. First, identify clock jumps through frequency sequence. Next, use threshold to get rid of abnormal data segments and combine median absolute deviation (MAD) and Baarda to detect and delete outliers. Then, replace missing points with linearly interpolated values. Finally, a clean SCB sequence is achieved. Based on the use of measured Beidou-2 SCB data，the test results show that the quality control strategy of clock bias can significantly improve the accuracy of real information extracted from Beidou-2 SCB and prediction.
