Transcranial magnetic stimulation (TMS) has been used to evaluate neuroplastic changes in the brain in clinical trials. The purpose of this study was to establish the test-retest reliability of 4 TMS measures of corticomotor excitability - (1) resting motor threshold, (2) slope of input-output curve, (3) peak motor evoked potential amplitude, and (4) cortical silent period duration for the corticospinal projections to the first dorsal interosseous of the contralateral hand. Fourteen healthy subjects (mean age 27.4 years) and 27 subjects with stroke-induced upper limb hemiparesis (mean age 61.3 years) completed 2 repeated sessions of assessment of 1 week apart. Good to excellent test-retest reliability of the TMS measurements was confirmed in the stroke subjects for both hemispheres with the ICC ≥ 0.88. Measurement reliability was good (ICC ≥ 0.75) for the 4 outcome measures in healthy subjects. Contrary to the similarity in standard error of measurements in both hemispheres for outcome measures (1) to (3) in the stroke subjects, that of the cortical silent period duration was larger in magnitude in the lesioned hemisphere. The test-retest reliability coefficients determined for the four corticomotor excitability measurements allowed the estimation of 95% minimal detectable changes of these outcome variables for the respective subject group in future clinical trials.