H-shaped steel columns are critical components in steel frame structures, necessitating performance-based seismic assessment. Fragility functions—cumulative distribution functions reflecting the probability of engineering demand parameters (EDP) exceeding damage thresholds—are pivotal but lack codified frameworks compliant with Chinese standards. This study integrates 94 global low-cycle reversed loading test datasets for H-shaped steel columns, adopting drift ratio as the EDP. Three damage states are classified based on structural degradation patterns, with tailored repair strategies. Fragility functions conforming to lognormal distribution are established using the FEMA P-58 methodology. The effects of axial compression ratio and cross-section classification are analyzed. Key findings: Elevated damage states significantly increase fragility function medians; Higher axial compression ratios accelerate plastic hinge formation and strength degradation, reducing medians; Reduced cross-section width-to-thickness ratios delay local buckling, enhancing medians; Cross-section classification outweighs axial compression ratio in influencing fragility curves. The proposed framework aligns with Chinese codes, addressing a research gap and supporting seismic resilience evaluation of steel structures.