Constraint-induced (CI) movement therapy is a promising therapy for improving upper limb function in adults after stroke. It involves restraint of the non-involved limb and extensive movement practice with the involved limb. In this study, a single-blinded, randomized, control study was performed to examine the efficacy of CI therapy, modified to be child friendly, in children with hemiplegic cerebral palsy (CP). Twenty-two children (8 females, 14 males; mean age 6y 8mo [SD 1y 4mo]; range 4–8y) were randomized to either an intervention group (n=11) or a delayed treatment control group (n=11). Children wore a sling on their non-involved upper limb for 6 hours per day for 10 out of 12 consecutive days and were engaged in play and functional activities. Children in the treatment group demonstrated improved movement efficiency and dexterity of the involved upper extremity, which were sustained through the 6-month evaluation period, as measured by the Jebsen–Taylor Test of Hand Function and fine motor-subtests of the Bruininks–Oseretsky Test of Motor Proficiency (p<0.05 in both cases). Initial severity of hand impairment and testing compliance were strong predictors of improvement. Caregivers reported significant increases in involved limb frequency of use and quality of movement. However, there was no change in strength, sensibility, or muscle tone (p>0.05 in all cases). Results suggest that for a carefully selected subgroup of children with hemiplegic CP, CI therapy modified to be child-friendly, appears to be efficacious in improving movement efficiency of the involved upper extremity.