In recent years, optimal control theory (OCT) has become a leading framework for studying the neural control of movement and motor cognition, contributing to two key research areas: behavioral neuroscience and humanoid robotics. Both fields face common challenges, such as the “degrees of freedom (DoFs) problem” and the fundamental processes of generating, observing, reasoning, and learning “actions.” Derived from control systems engineering, OCT quantifies task goals through “cost functions” and employs advanced mathematical tools to achieve desired behaviors and make predictions. On the other hand, action selection is a critical decision-making process that relies on assessing the state of both the body and the environment. Given that sensory and motor signals are often affected by variability and noise, the nervous system must estimate these states. To choose the optimal action, these state estimates must be integrated with knowledge of the potential costs or rewards associated with different outcomes. This paper reviews recent research on how the nervous system addresses these estimation and decision-making challenges, highlighting findings that suggest human behavior closely aligns with predictions from Bayesian Decision Theory. This theory defines optimal behavior in uncertain environments and offers a coherent framework for understanding sensorimotor processes.