Background Equivalent continuous A-weighted sound pressure level is not appropriate for evaluating the risk of non-steady noise exposure, and need to be corrected by noise time-domain structure, but the correction method and its applicability need to be discussed. Objective To validate the application of the kurtosis-adjusted normalization of equivalent continuous A-weighted sound pressure level to a normal 8 h working day ( L_{Aeq,8 h}) in assessing noise-induced hearing loss (NIHL), and to improve the methods for assessing occupational hearing loss associated with different types of noise. Methods Audiometric and shift-long noise exposure data were acquired from a population(n=2 466) of screened workers exposed to noise between 70 dB(A) and 95 dB(A) from 6 industries in China. The cohort data were collapsed into 1 dB(A) bins, and the average kurtosis and noise-induced permanent threshold shifts at 3 kHz, 4 kHz, and 6 kHz (NIPTS₃₄₆) within 1 dB(A) were calculated respectively. According to the existing correction method, the adjustment coefficient λ was calculated by multiple regression, and L_{Aeq,8 h} was corrected by λ (L'_{Aeq,8 h}). The entire cohort was divided into K₁ (≤10; steady noise), K₂ (10~50; non-steady noise), and K₃ (＞50; non-steady noise) groups based on mean kurtosis levels. Predicted NIPTS₃₄₆ was calculated using the ISO 1999 model for each participant and the actual measured NIPTS₃₄₆ was corrected for age and gender. The underestimated NIPTS₃₄₆ was the difference between the values of estimated NIPTS₃₄₆ and the corresponding actual NIPTS₃₄₆. To validate the applicability of L′_{Aeq,8 h} in evaluating NIHL, the correlation between L′_{Aeq,8 h} and HFNIHL, and the mean difference between real NIPTS₃₄₆ and estimated NIPTS₃₄₆ were analyzed. Results The adjustment coefficient λ was determined at 5.43. The results of multiple logistic regression analysis showed that the relationship between L'_{Aeq,8 h} and HFNIHL increased from 6.6% to 9.6% after the kurtosis adjustment. The DRR of L_{Aeq,8 h} and HFNIHL showed that the percentage of HFNIHL decreased after the adjustment of kurtosis in the non-steady noise groups, and the regression lines of the non-steady noise groups approached that of the steady noise group. The R² of the K₂ group increased from 0.935 3 to 0.986 3, and the R ² of the K₃ group increased from 0.905 6 to 0.951 6. Under the un-adjusted condition, the NIPTS₃₄₆ underestimation for the K₃ group was significantly higher than that for the steady noise group (t=−3.23, P=0.001). After the L_{Aeq,8 h} was adjusted by kurtosis, the NIPTS₃₄₆ underestimation decreased significantly in the three kurtosis groups (K₁: t=6.78, P＜0.001; K₂: t=14.31, P＜0.001; K₃: t=11.06, P＜0.001). There was no significant difference in the degree of underestimation between the three kurtosis groups (K₁ vs K₂: t=−0.22, P=0.830; K₁ vs K₃: t=−1.40, P=0.205) as the curves of the three kurtosis groups were nearly overlapped. Conclusion The kurtosis-adjusted L_{Aeq,8 h} can effectively estimate the hearing loss associated with non-steady state noise.