Abstract: In order to improve the positioning and timing accuracy of BDS, it is necessary to calculate the satellite clock difference accurately. Considering the influence of random part error in clock difference of BDS, the prediction model of clock difference is established by fitting estimation. The model is solved by two-step minimization method based on fitting estimation. In solving covariance function coefficients, the polynomial fitting coefficients are fitted by model residual. Finally, the results of the model are compared with those of the grey theory model and the normal method of fitting and estimating. The experimental results show that the model is effective and applicable in the short-term prediction and interpolation of the clock difference of BDS.

Key words: BDS, grey theory, fitting, cubic spline interpolation, two step’s minimization