<p>Accurate estimation of gravity-height gradient and plumb line curvature is fundamental for gravity field modeling, geoid determination, and high-precision geodetic applications, especially in mountainous regions. This study presents an integrated computational framework combining terrestrial gravity observations, precise leveling, GNSS measurements, and Global Geopotential Models (GGMs) to estimate and compare gravity-height gradient and plumb line curvature in the rugged terrain surrounding Sümela Monastery. Simultaneous gravity and leveling surveys were conducted during two observation periods to compute actual and normal gradient, level surface curvature, and plumb line curvature. Additionally, twelve GGMs obtained from the ICGEM service were used to derive the same parameters through spherical harmonic expansions. A comprehensive statistical evaluation, including minimum, maximum, mean, standard deviation, and RMSE, was performed to assess the consistency between terrestrial and model-based results. The results show that normal gradient and level surface curvature from GGMs generally agree well with terrestrial data, whereas larger discrepancies occur in actual gradient and plumb line curvature in areas with steep topographic gradients. CHAMP-based models demonstrate comparatively better consistency in representing plumb line curvature. The proposed framework confirms the reliability of global models at regional scales while emphasizing the importance of terrestrial data in complex topographic environments.</p>