Objective: To determine whether functional electrical stimulation (FES) timed to activate the dorsiflexors and plantar flexors during gait improves the walking of adults with hemiplegia. Design: Randomized crossover trial. Setting: Outpatient rehabilitation clinic. Participants: Adults with hemiplegia (N=28) with a mean age +/- SD of 60 +/- 10.9 years and 4.9 +/- 3.8 years postincident. Interventions: Intervention "A" included 3 months of wearing the FES system, which activated automatically during walking for 6 to 8h/d, 7d/wk, plus walking 1h/d, 6d/wk. Intervention "B" included 3 months of walking 1 h/d, 6d/wk without FES. Of the 28 patients who completed the study, 15 were randomly assigned to group A-B, 13 to group B-A. Crossover occurred at 3 months. Main Outcome Measures: Variables were measured at pretreatment, 3 months, and 6 months. Three primary outcomes were selected a priori and included 2 functional variables, the 6-minute walk test and the Emory Functional Ambulatory Profile, and I participation variable, the Stroke Impact Scale. Secondary impairment measures included muscle strength and spasticity. Assessments were done without electrical stimulation. Results: In phase 1, patients who received treatment A (A-B group) showed improvement compared with patients who received treatment B (B-A group) on the 6-minute walk test (P=.02), Emory Functional Ambulatory Profile (P=.08), and Stroke Impact Scale (P=.03). In phase 2, the A-B group maintained improvement in all 3 primary outcomes even without FES. Both groups improved significantly on all primary outcome measures, comparing 6-month to initial measures (P <=.05). Conclusions: An FES system that stimulates dorsiflexors and plantar flexors similar to the timing of typical adult gait, combined with daily walking, can improve the walking ability of adults with hemiplegia.