Objective: To examine whether alternating training with both the non-paretic and paretic sides (alternating bilateral training), expecting trial-to-trial inter-limb transfer of training effects from the nonparetic to the paretic side, improves upper-limb motor performance in post-stroke patients, compared with unilateral training involving only the paretic side.Design: An assessor-blinded pilot randomized controlled trial.Subjects: Twenty-four right-handed post-stroke patients with hemiparesis.Methods: Participants were randomly allocated to either an alternating bilateral training group or a unilateral training group (n = 12/group). Participants underwent dexterity training of the paretic arm using the Nine-Hole Peg Test, completing 10 trials/day for 7 consecutive days. The alternating bilateral training group additionally performed alternating trials with the non-paretic limb. Performance change, assessed 1 day and 1 week after the 7-day training period, was compared between groups.Results: Although the improvement was comparable in both groups at both post-training time-points, a sub-analysis in which those with left hemiparesis and those with right hemiparesis were analysed separately revealed potential benefits of the alternating bilateral training, specifically for those with left hemiparesis.Conclusion: Alternating bilateral training may augment training effects and improve upper-limb motor function in patients with left hemiparesis.