Future planetary explorations require versatile robots that can adaptively traverse extreme access and partially known environments with optimal energy-consumption to address the limitations of current exploration rovers and robotic systems, like Mars and Lunar Exploration Rovers. This presentation discusses a study of the dynamics, modeling, and dynamic-based optimization for energy-efficient agile snake robot locomotion for space exploration. Muscle-driven locomotion for a planar snake robot will be introduced which integrates the advantages of flexibility and robustness that come from both soft and rigid robotics, respectively. The system is comprised of two adjacent links connected by a pair of antagonistic pneumatic artificial muscles (PAMs) each coupled with an extension spring. An alternate actuation of these soft actuators causes rotational motion about the connecting joint. When the PAM is passive, it has little to no capability to stretch, limiting its range of motion. To address this issue extension springs were implemented in this design to compensate for the lack of needed stretch. Such a combination of rigid and soft robotic approaches is essential when prioritizing the versatility, adaptability, and optimal performance capabilities that are vital for a terrestrial space robot. Kinematics of the muscle-driven robot in the joint and Cartesian space were derived with respect to the motion of the muscles. Lagrangian mechanics was utilized for obtaining the dynamic model of the robot as well as to derive the system's equations of motion. The performance of the dynamic model was then demonstrated through simulation using MATLAB for a three-link muscle-driven snake-like robot. The dynamic modeling for an N-link system will be discussed. Following the modeling of the dynamics for the system, a velocity-based optimization was carried out on the dynamic model to analyze the optimal values of the geometric parameters of the linkage and dynamical properties of the model that would yield optimal forward velocity performance. This research is based upon work supported by the New Mexico Space Grant Consortium (NMSGC) Space Grant Fellowship through a NASA Cooperative Agreement No. NM-80NSSC20M0034.