Background: The purpose of the study is to investigate the correlation between upper lumbar disc herniation (ULDH) and multifidus muscle degeneration via the comparison of width, the cross-sectional area and degree of fatty infiltration of the lumbar multifidus muscle. Methods: Using the axial T2-weighted images of magnetic resonance imaging as an assessment tool, we retrospectively investigated 132 patients with ULDH and 132 healthy individuals. The total muscle cross-sectional area (TMCSA) and the pure muscle cross-sectional area (PMCSA) of the multifidus muscle at the L1/2, L2/3, and L3/4 intervertebral disc levels were measured respectively, and in the meantime, the average multifidus muscle width (AMMW) and degree of fatty infiltration of bilateral multifidus muscle were evaluated. The resulting data were analyzed to determine the presence/absence of statistical significance between the study and control groups. Multivariate logistical regression analyses were used to evaluate the correlation between ULDH and multifidus degeneration. Results: The results of the analysis of the two groups showed that there were statistically significant differences (p < 0.05) between TMCSA, PMCSA, AMMW and degree of fatty infiltration. The multivariate logistic regression analysis indicated that the TMCSA, PMCSA, AMMW and the degree of fatty infiltration of multifidus muscle were correlated with ULDH, and the differences were statistically significant (P < 0.05). Conclusions: A correlation could exist between multifidus muscles degeneration and ULDH, that may be a process of mutual influence and interaction. Lumbar muscle strengthening training could prevent and improve muscle atrophy and degeneration. © 2021, The Author(s).