The capacity to reacquire motor skills lost after a stroke is crucial to promote upper-limb motor recovery, but the impact of lesion location on motor skill acquisition and the underlying neurophysiological mechanisms remain uncertain. Cortico-spinal excitability, measured with transcranial magnetic stimulation, provides information about the functional integrity of the cortico-spinal tract, a structure essential for upper-limb motor function and recovery post-stroke. Cortico-spinal excitability is a biomarker that has been used to predict upper-limb motor recovery and guide interventions post-stroke. We investigated associations between different cortico-spinal excitability measures of facilitation and inhibition and the capacity to improve the performance of a short-term visuomotor adaptation task that required accurate force modulation with the most affected hand in 103 individuals with either cortical (n = 34) or subcortical (n = 69) lesions. Cortico-spinal excitability measures included resting motor threshold, resting and active excitability derived from the amplitude of motor evoked potentials, cortical silent period, as well as intracortical facilitation and short intracortical inhibition. Both lesion groups showed similar rates of motor skill performance but individuals with subcortical lesions exhibited more impairment in the most affected hand and lower excitability in the ipsilesional hemisphere inferred from a reduced amplitude of motor evoked potentials elicited during an active muscle contraction. Exploratory analyses revealed that upper-limb impairments and reductions in active and resting excitability in the ipsilesional hemisphere were exacerbated in individuals with subcortical lesions affecting the cortico-spinal tract. In individuals with cortical lesions, better motor skill performance was associated with lower motor thresholds (β = -0.25, 95% CI [-0.47, -0.03]; P = 0.024) and higher intracortical inhibition (β = -3.93, 95% CI [-6.89, -0.98]; P = 0.011) in the ipsilesional hemisphere. In contrast, in individuals with subcortical lesions motor skill performance was associated with smaller motor evoked potentials (β = -4.46, 95% CI [-8.54, -0.38]; P = 0.033), less intracortical inhibition (β = 3.45, 95% CI [0.34,6.56]; P = 0.030) and higher facilitation (β = 1.34, 95% CI [0.15,2.54]; P = 0.028) in the ipsilesional hemisphere. Associations with intracortical inhibition and facilitation in the subcortical group were driven primarily by observations from individuals with lesions affecting the cortico-spinal tract. Importantly, no associations were found in the contralesional hemisphere. Reinforcing the existence of lesion-specific neurophysiological reorganization patterns, individuals with cortical and subcortical lesions show divergent associations between cortico-spinal excitability and the capacity to improve the performance of a short-term visuomotor adaptation task. The use of cortico-spinal excitability as a biomarker to guide motor recovery interventions such as non-invasive brain stimulation or to predict upper-limb recovery post-stroke should consider lesion location.