OBJECTIVE: To investigate the within-session reliability and test-re-test reliability of hand-held dynamometry (HHD) to quantify strength in hip flexors, quadriceps, plantarflexors and triceps brachii following traumatic brain injury (TBI). DESIGN: Within-session reliability was examined across six consecutive trials of HHD. Test-re-test reliability was examined between two sessions separated by 7 days. SETTING: Private physical therapy practice, Melbourne, Australia. PARTICIPANTS: Ten adults with TBI with a mean age of 32.5 years (range = 19-63, SD = 12.8) and mean time after injury of 12.2 years (range = 2.3-26, SD = 8.6). MAIN OUTCOME MEASURES: Isometric strength measured by HHD. RESULTS: Test-re-test reliability of HHD following TBI was good when repeated tests occurred within a single session. Within a session, the first trial was the most variable out of the six examined. Of the four muscles examined, re-test measures of strength for the quadriceps, plantarflexors and triceps were the most reliable. For the hip flexors an increase greater than 3.1 kg (28%) was needed before real change could be detected. Test-re-test reliability was higher for muscles tested on the more affected side (ICC((2,1)) = 0.55-0.93) than muscles tested on the less affected side (ICC((2,1)) = 0.09-0.86). CONCLUSIONS: To gain a representative measure of isometric strength using HHD in a single session, three trials are recommended. The first is for familiarization and trials 2 and 3 are averaged to provide a typical measure of isometric muscle strength. Further research needs to occur but alternative methods to quantify muscle strength might be needed if repeated measures are to be performed over a 7-day interval.