Abstract: Collapse assessment requires component hysteretic model that accurately captures deterioration behaviors under different loading paths. The existing path-dependent RC column hysteretic model relies on parameter calibration derived predominantly from experimental data with high-strength concrete, high axial compression ratios (ACR), and high reinforcement ratios. These data mismatch fail to represent critical parameter characteristics of low- to mid-rise RC frame structures, limiting applicability. To enhance collapse assessment accuracy for such structures, this study systematically evaluates existing models’ accuracy in simulating mechanical behavior of such structures. An RC column experimental database reflecting key structural characteristics (low-strength, low ACR, low reinforcement ratios) was established. Using the dual metrics of prediction mean and uncertainty, and combining physical mechanism interpretation with statistical analysis, the predictive accuracy of existing model parameter formulas were evaluated based on established database. Results demonstrate that elastic parameter formulas exhibit lower uncertainty and higher accuracy, while nonlinear formulas show significant underestimation. Consequently, error-prone and missing nonlinear formulas were improved and supplemented, yielding the improved RC column hysteretic model. Experimental validation confirms the improved model adequately incorporates low-strength, low-ACR, low-reinforcement design features, enabling more accurate simulation of mechanical behavior throughout the seismic collapse process, from elastic response to deterioration and failure.