This study presents an integrated geotechnical and site-specific nonlinear ground response assessment of Islamabad, Pakistan, to evaluate the influence of local soil conditions on seismic ground-motion amplification. The methodology combined field investigation, laboratory index testing, resonant column and cyclic triaxial testing, shear-wave velocity characterization, and one-dimensional nonlinear site response analysis. Thirty boreholes were investigated across Islamabad, and recovered samples were tested for index and dynamic soil properties, including grain-size distribution, Atterberg limits, moisture content, and specific gravity. The subsurface soils are predominantly low-plasticity silts and silty clays, mainly classified as ML and CL-ML, indicating fine-grained alluvial deposits with low to moderate plasticity. Resonant column tests defined the small-strain stiffness and damping response, whereas cyclic triaxial tests extended the characterization to medium and large strain levels. The two datasets were integrated to develop continuous laboratory-calibrated modulus reduction and damping curves for nonlinear modeling. Ten representative soil profiles were developed using borehole logs, unit weights, index properties, and locally calibrated SPT-based shear-wave velocity estimates. Based on Vs30, profiles S1 to S8 were classified as BCP 2021 Site Class C, whereas S9 and S10 were classified as Site Class D. The profiles were analyzed in DEEPSOIL v7.0 using the GQ/H nonlinear constitutive model, with laboratory-calibrated curves used as the principal input for experimental nonlinear response analysis. The nonlinear analyses indicate dominant short-period amplification within approximately 0.1-0.4 s. Spectral amplification factors generally range from about 1.5 to 2.3, with the strongest amplification occurring in intermediate-depth Site Class C profiles where impedance contrast and strain-compatible stiffness promote resonance. Nonlinear deformation is concentrated mainly in the upper 5 to 10 m of the profiles, whereas deeper Site Class D profiles distribute strain over greater thickness but do not show a sustained long-period spectral plateau. The laboratory-calibrated results further indicate that stiffness degradation becomes pronounced beyond about 0.1% shear strain, while damping increases to approximately 10 to 15% at larger strains, thereby limiting excessive amplification through hysteretic energy dissipation. Overall, the findings demonstrate that experimentally constrained nonlinear site response analysis is essential for realistic seismic demand assessment in Islamabad and that Vs30-based classification alone may not adequately capture the effects of local stratigraphy, impedance contrast, and measured nonlinear soil behavior.