We examined the effects of previous trials on subsequent trials on performance in the continuous performance test (CPT) in children with attention-deficit-hyperactivity disorder (ADHD). Thirty-five non-medicated children with ADHD (31 males, four females; mean age 9y 10mo [SD 2y 4mo]) and 33 comparison children (20 males, 13 females; mean age 10y [SD 2y 7mo]) were tested using a novel CPT, in which stimuli were presented with 50% target probability. Reaction time, reaction time variability, omission, and commission error rate were analyzed for two different types of trials in which different responses (switched trials) or the same responses (repeated trials) were required for two consecutive trials. Compared with the comparison group, children with ADHD showed a greater increase in commission error rate from repeated to switched trials, i.e. increased switch cost for commission error rate. On the other hand, omission error rate was not influenced by the previous trial in both ADHD and comparison groups. These results suggest that children with ADHD have trouble in response switching when an inhibitory process is involved.