The determination of orthometric heights with high precision is vital for geospatial analysis, engineering design, and modern infrastructure planning. While spirit levelling technique is accurate, it is labor-intensive and less practical for large areas or difficult terrain. This study adopts a GNSS-based approach at Abiola Ajimobi Technical University (AATU), integrating multiple global geoid models to enhance vertical accuracy. A dual-frequency DGPS receiver, linked to a CORS network, was used to obtain ellipsoidal heights from 300 stochastically and randomly distributed points. Five gravimetric geoid models; EGM2008, EIGEN-6C4, GECO, SGG-UGM-2, and XGM2019e, were applied to compute geoid undulations and convert ellipsoidal heights to orthometric heights. Data processing involved QGIS for spatial mapping, statistical software for accuracy checks, and online geoid computation tools for verification. Quality control steps included outlier detection, cross-model comparisons, and heatmap analysis of deviations. Results showed that EGM2008 and XGM2019e had the closest agreement, while SGG-UGM 2 recorded larger discrepancies, highlighting the importance of model selection for local geophysical conditions. The workflow demonstrates that GNSS with appropriate geoid models can reliably deliver engineering grade orthometric heights, offering a replicable, efficient, and precise method for infrastructure development and geodetic modernization in developing regions.