Abstract: ADHD is characterized by a difficulty to act in a goal directed manner. Even though most tasks in natural environments require a sequence of actions to be acted in order to obtain a goal, ADHD was never studied in the context of value-based sequence learning. Here, we made use of current advancements in hierarchical reinforcement learning algorithms to track the internal value and choice policy of individuals with ADHD performing a three-stage sequence learning task. Specifically, 54 participants (28 ADHD, 26 TD) completed a well-established value-based reinforcement learning task that allowed us to estimate latent internal action-values for each trial and stage using computational modeling. We found attenuated sensitivity to action values in ADHD compared with their TD comparison peers, both in choice and reaction-time variability estimates. Remarkably, this was found only for first stage actions (i.e., initiatory actions), while for action performed just before outcome delivery (i.e., terminal actions) the two groups were strikingly indistinguishable. Overall, these results suggest a difficulty in value estimation for initiatory actions in ADHD. We discuss the implication of our findings to the ability to initiate tasks and follow them to their end goal in ADHD.