The effects of interface defect states with different distribution forms on the C-V characteristics and interface charge modulation behavior of Al₂O₃/Y₂O₃/In_0.53Ga_0.47As MOS structures were investigated. Experimental data and TCAD simulations were combined to analyze their underlying mechanisms. Using the experimental C-V curves at 100 Hz and 1 MHz as calibration benchmarks, a comprehensive interface defect state model consisting of near-band-edge exponentially distributed continuous defect states and discrete energy level defect states was established at the InGaAs/Y₂O₃ interface. By adjusting the defect state type, we analyzed the reference energy level, energy position, defect state density, and characteristic energy parameter, the response characteristics of different defect states. Based on the evolution of band bending, carrier distribution, and interface trapped charge variation during gate-voltage sweeping, the formation mechanisms of the overall stretch-out and local anomalous fluctuations in the experimental C-V curves were further clarified. The results show that discrete donor-like and acceptor-like interface defect states mainly produce local responses when the Fermi level sweeps across specific energy levels. Through enhanced electric-field screening and the introduction of defect capacitance, they can induce local capacitance reduction or hump-like behavior, accompanied by Fermi-level pinning. In contrast, exponentially distributed interface defect states mainly modulate the interface trapped charge and the associated screening effect through charging and discharging of defect state over a wide energy range, thereby affecting the overall stretch-out of the C-V curves over a wide bias range. Among them, exponentially distributed donor-like defect states near the valence-band side mainly modulate the response in the accumulation region, whereas exponentially distributed acceptor-like defect states near the conduction-band side mainly suppress inversion electron accumulation. These results indicate that a single defect state distribution model is insufficient to simultaneously explain the overall profile and local details of the experimental curves, and that the synergistic effects of continuous and discrete defect states must be considered together. The comprehensive interface defect state model and analysis method established in this work provides a useful framework for studies on the modulation mechanisms of interface defect states, parameter extraction, and interface process optimization in InGaAs MOS structures.