Owing to variations in data sources, inversion methods, and data processing strategies in global geopotential models (GGMs) derived by different institutions, there exists divergence in precision across different regions. Consequently, integrating information from various GGMs holds significant practical value for deriving a more robust GGM with consistent precision. This study combines GGMs such as XGM2019e_2159 and EIGEN- 6C4 through arithmetic mean, weighted mean aggregation, and variance component estimation (VCE) techniques, with respecting to a reference model. Moreover, the evaluation encompasses GNSS leveling assessments alongside DTU17 marine gravitational anomalies and regional geoid modeling analyses. The results of the comparative analysis demonstrate that: (1) Whether to rescale with respect to the reference model has a relatively minor impact, with a maximum impact of 1.4 mm on height anomalies while affecting gravimetric anomalies by 0.05 mGal; (2) VCE (Scheme a) and arithmetic mean methods yield comparatively stable GGMs showcasing strong performance across all regions according to external validation from multiple sources.