In this paper, a ground-penetrating radar (GPR) system for the detection and localization of buried human bones in battlefield conditions is numerically modeled and simulated. The finite-difference time-domain (FDTD) method with the complex perfectly matched layer (CPML) absorbing boundary condition is employed to simulate wave propagation and interaction with the soil and target. The electromagnetic parameters of the soil (based on real samples from the Shalamcheh and Tala'eyeh regions in Khuzestan Province) and bone (measured experimentally on fresh and aged specimens in the frequency range of 200 MHz to 2 GHz) are incorporated into the model. Quantitative results show that at 250 MHz and a depth of 2 m, the signal-to-noise ratio (SNR) is 14.86 dB, while at 4 m depth (200 MHz) the SNR is 2.37 dB. The presence of a metallic object (helmet) near the bone nonlinearly increases the SNR by up to 31 dB. Lower frequencies (200–250 MHz) are more suitable for deeper penetration, whereas higher frequencies (500 MHz) provide better resolution for small targets. A white Gaussian noise model is assumed, and a detection threshold of approximately 12 dB is determined. The probability of detection (Pd) at 2 m depth exceeds 0.95, and the probability of false alarm (Pfa) is estimated to be below 0.01. Due to the limitations of the 2D model and the assumption of homogeneous soil, the results should be interpreted with caution when applied to real field conditions