Does injection of Botulinum toxin followed by constraint-induced movement therapy improve spasticity and upper limb motor function more than the same injection followed by rehabilitation based on neurodevelopmental techniques? Randomised trial with concealed allocation and blinded outcome assessment. Rehabilitation department of a tertiary hospital in Taiwan. Adults at least one year after a stroke with a Modified Ashworth Scale (MAS) score of 3 or more in the elbow, wrist, or finger flexors, and with at least 10 degrees of active interphalangeal and metacarpophalangeal extension and 20 degrees of wrist extension. Fixed contractures, major co-morbidities, and previous Botulinum toxin injection or surgery for spasticity were exclusion criteria. Randomisation of 32 participants allotted 16 to each group. Both groups received a total dose of 1000 units of Botulinum toxin type A, injected at standard muscular sites in the affected upper limb, and commenced their 3-month rehabilitation regimen the following day. The intervention group underwent intensive training of the affected upper limb for 2 hours, 3 times per week, while the non-affected upper limb was restrained for at least 5 hours per day. Selected tasks were progressed in complexity, with some assistance with movements and verbal feedback and encouragement. The control group received 1 hour each of physiotherapy and occupational therapy, 3 times per week. Therapy was based on neurodevelopmental techniques, focusing on normalising tone, and movement patterns. The primary outcome was the MAS (0 = no spasticity, 4 = rigid in flexion or extension). Secondary outcomes included the Motor Activity Log (MAL), comprising two 6-point scales of amount of use and quality of movement, and the Action Research Arm Test (ARAT), which rates 19 tasks from 0 (no movement possible) to 3 (normal movement), to give a total score out of 57. 29 participants completed the study. At 6 months, the treatment group had significantly greater reduction in MAS scores for the elbow (0.7, 95% CI 0.1 to 1.3), wrist (0.7, 95% CI 0.2 to 1.2), and fingers (1.2, 95% CI 0.9 to 1.5). Also at 6 months, the treatment group had significantly greater improvement in amount of use (1.1, 95% CI 0.8 to 1.4), quality of movement (0.9, 95% CI 0.6 to 1.2), and ARAT scores (7, 95% CI 4 to 10). Injection of Botulinum toxin followed by constraint-induced movement therapy improves spasticity and upper limb motor function more than the same injection followed by rehabilitation based on neurodevelopmental techniques.