Purpose: To investigate whether pre-operative magnetic resonance imaging (MRI) of the lumbar multifidus muscle (LMM) would predict clinical outcomes following lumbar spinal decompression for symptomatic spinal stenosis. Methods: A prospective cohort of patients with symptomatic neurogenic claudication, documented spinal stenosis on pre-operative MRI underwent spinal decompression. All subjects completed standardised outcome measures (Core Outcome Measures Index (COMI), Oswestry Disability Index (ODI v2.1) pre-operatively, 1 and 2 years post-surgery. Surgery was performed using a standardised lumbar spinous process osteotomy for access, followed by a decompression of the central canal, lateral recess and foraminal zones as indicated by the pre-operative MRI. Lumbar MRI scans were evaluated by two independent observers who assessed the axial CSA of the LMM bilaterally and the degree of muscle atrophy according to the Kader classification (2000). Changes in COMI and ODI scores at 1 and 2 years were investigated for statistically significant correlations with CSA of LMM and Kader grading. Statistical analyses utilised Student’s t test, kappa coefficient for inter-observer agreement and Bland–Altman Limits of Agreement (BALOA). Results: 66 patients (41 female) aged between 29 and 86 years underwent single-level decompression in 44, two-level decompression in 16 and three-level decompression in 6 cases. No significant correlation was observed between improvements in ODI and COMI relative to age, degree of stenosis, posterior fat thickness or psoas CSA. Those subjects with the greatest LMM atrophy relative to psoas CSA and L5 vertebral body area on pre-operative MRI had the least absolute improvement in both ODI and COMI scores (p = 0.006). Conclusions: Reduced LMM CSA (<8.5 cm2) and muscle atrophy were associated with less favourable outcomes following lumbar spinal decompression. Pre-operative CSA of LMM appeared to be a more reliable predictor of post-operative clinical outcomes compared to the Kader Grading Score. This is the first study to investigate the prognostic value of pre-operative MRI appearance and CSA of LMM with respect to post-operative outcome following lumbar decompression surgery. Healthy pre-operative LMM is associated with better outcomes following lumbar spinal decompression.