Abstract:GNSS triple-frequency is advantageous in reducing observation errors, so this paper advances a new algorithm of fast carrier phase observation accuracy estimation. Based on the assumption that GNSS triple-frequency carrier observations are independent, we use sliding polynomial fitting to get the observational precision of triple-frequency carrier GIF(Geometry-Free Ionosphere-Free) combinations. GPS and BDS triple-frequency experimental data results show that the algorithm can be used to quickly estimate the triple-frequency carrier phase observation precision, so it can provide a fast and accurate random function model for single precise point positioning.