A global navigation satellite system (GNSS) receiver with high sensitivity is embarked by the Chang’E-5T spacecraft for a trial in the Chinese lunar exploration program. The GNSS flight experiment was activated twice on October 23 and 31, 2014. The measurement data of this experiment are analyzed here. When the distance between the Earth center and Chang’E-5T probe ranges from approximately 10,000–60,000km, the mean number of tracked GNSS satellites by the receiver onboard Chang’E-5T is 8–9, and the range of Position Dilution of Precision (PDOP) is 1–40. The noise of single-differencing C/A code pseudorange data is 5.7–8.1m (1σ). The ratio of carrier to noise spectral density (C/N0) is approximately 35dB-Hz when the distance between the spacecraft and GPS satellites was close to 60,000km. The post-processed orbit solution using terrestrial-based radiometric measurement in the 22h arc is used as the reference orbit for accuracy analysis of orbit determination and prediction. Based on this approach, the position error in the one-hour prediction from the orbit determination using GNSS single-difference pseudorange in the 1.5h arc is less than 109m, while that from the result using range and Doppler data in the 3h arc is 369m. The residual Root Mean Squares (RMS) divergence of GNSS single differential pseudorange and range in the prediction arc relative to the orbit determination arc are reduced by 65% and 90%, respectively, using GNSS data to estimate the orbit compared to using ground-based data. Furthermore, the strategy of orbit determination by combining GNSS and ground-based data is examined in the paper. The position error in the one-hour prediction using this strategy is less than 47m using the 1.5h data to determine the orbit, and the residual RMS divergences of GNSS and range measurement are reduced by 96% and 67%, respectively. Therefore, the validity of GNSS supporting orbit determination for sampling and returning from lunar explorations is verified.