The paper describes the design and analysis of Radioisotope Thermoelectric Generators (RTGs) for powering the Mars Rover vehicle, which is a critical element of the Mars Rover and Sample Return mission (MRSR).The RTG design study was conducted by Fairchild Space Company for the Department of Energy, in support of the Jet Propulsion Laboratory's MRSR project. The paper briefly describes a reference mission scenario, an illustrative Rover design and activity pattern on Mars, and its power system requirements and environmental constraints, including the RTG cooling requirements during transit to Mars. It identifies the key RTG design problem, i.e. venting the helium generated by the fuel's alpha decay without intrusion of the Martian atmosphere into the RTG, and proposes a design approach for solving that problem. Using that approach, it describes and analyzes a variety of RTG designs, all based on the proven and safetyqualified General Purpose Heat Source Module. The first RTG option described is a very conservative baseline design, employing standard thermoelectric unicouples whose reliability and performance stability has been extensively demonstrated on previous space missions. The heat source of the 250-watt RTG consists of a stack of 18 separate modules that is supported at its ends but not along its length. The paper describes and analyzes the structure that holds the stack together during Earth launch and Mars operations, but allows it to come apart in case of an inadvertent reentry. It then summarizes the baseline RTG's mass breakdown, and presents a detailed description of its thermal, thermoelectric, and electrical analyses. It examines the effect of different operating conditions (BOM/EOM, watercooled/radiation-cooled, summer-day/winter-night) on the RTG's performance. Finally, the paper compares the RTGs' specific powers for different power levels (250W/125W), different thermoelectric element designs (standard unicouples/short unicouples/multicouples), and different thermoelectric figures of merit (0.00058K ). The results presented show the RTG performance achievable with current technology, and the performance improvements that would be achievable with various technology developments. It provides a basis for selecting the optimum strategy for meeting the Mars Rover design goals with minimal programmatic risk and cost.