Abstract: The BeiDou Navigation Satellite System (BDS-3) officially began providing global services on July 31, 2020. As the core component of satellite navigation systems, the space segment's service capability determines the overall system performance. This paper first introduces calculation and analysis methods for broadcast orbit errors, broadcast clock errors, Signal-in-space range error and broadcast ionospheric errors. Then, based on the final products of German Geosciences Research Centre (GFZ)and International Global Navigation Satellite System Monitoring and Assessment System (iGMAS),it conducts an evaluation of the changes in broadcast ephemeris accuracy and ionospheric model precision throughout a complete cycle from 2020 to 2025. Research shows that the BDS‑3 broadcast orbit error shows a clear dependence on satellite type, with MEO satellites outperforming IGSO satellites. Compared with GFZ products ,the 95% RMS of the radial ,along‑track and cross‑track error falls from 0.104 m、0.482m、0.589m respectively in 2020 to 0.080 m、0.351m、0.364 m respectively in 2025; and compared with iGMAS products , these data fall from 0.086m、0.386m、0.461m respectively  in 2020 to 0.073m、0.341m、0.350m in 2025.Regarding broadcast clock error, based on GFZ products,the 95% RMS improves from 0.705m in 2020 to 0.540 m in 2025; and based on iGMAS products, the 95% RMS in 2020 is 0.811m and 0.640m in 2025.The SISRE of MEO satellites is generally smaller than that of IGSO satellites, and the statistical accuracies of 95% RMS of SISRE in 2020 based on GFZ and iGMAS products reaches 0.705m and 0.817m,and in 2025 reaches 0.549m and 0.645m, respectively. In terms of ionospheric model error, during the solar minimum(2020) the BDGIM model’s VTEC average RMS is 3.193 TECU for the Center for Orbit Determination in Europe(CODE), and 3.176TECU for iGMAS(vs. 7.359 TECU for the Klobuchar model, iGMAS for 7.367TECU), while during the solar maximum(2025) they rise to 11.481 TECU and 22.211 TECU for CODE, 10.493TECU and 21.802TECU for iGMAS respectively, demonstrating BDGIM model’s stronger resistance to disturbances. And the assessment results can provide support for BDS performance optimization.