Training can improve human decision-making performance. After several training sessions, a person can quickly and accurately complete a task. However, decision-making is always a trade-off between accuracy and response time. Factors such as age and drug abuse can affect the decision-making process. This study examines how training can improve the performance of different age groups in completing a random dot motion (RDM) task. The participants are divided into two groups: old and young. They undergo a three-phase training and then repeat the same RDM task. The hierarchical drift-diffusion model analyzes the subjects' responses and determines how the model's parameters change after training for both age groups. The results show that after training, the participants were able to accumulate sensory information faster, and the model drift rate increased. However, their decision boundary decreased as they became more confident and had a lower decision-making threshold. Additionally, the old group had a higher boundary and lower drift rate in both pre and post-training, and there was less difference between the two group parameters after training.