Planetary exploration rovers are required to (semi-) autonomously navigate itself on rough terrain. In such scenario, a feasible path to be followed must be accurately planned, allowing the rover to safely traverse without any mobility hazards. In this paper, a path planning and evaluation strategy that explicitly considers dynamic mobility of the rover is presented. The proposed strategy consists of the following three steps. First, various paths on a given terrain map are generated with varying weighting factors for path planning algorithm. Each path is then examined through complete dynamic simulations of the rover in which the rover travels along with the path. The simulation result provides a metric for robotic mobility. The metric, termed a dynamic mobility index, consists of stability of the rover, wheel slippage, elapsed time, and energy consumption. All of the paths generated are quantitatively evaluated based on the dynamic mobility index, and then, the most feasible path between them is obtained. Demonstrations for the path planning and evaluation are presented in this paper that confirms the validity of the proposed strategy.