By employing concepts such as “option” and “strategy” from game theory, this study derives an ecologically-oriented dendrogram of the probable evolutionary history and the present relationships of sea turtles. An “armored tank” strategy is seen as differentiating the first ancestral testudines from the stem reptiles and providing enduring advantages while simultaneously imposing basic restrictions on all later forms. A “flipper” strategy is postulated as basic to development of the sea turtle line, again imposing limitations while conferring selective advantage. Modern sea turtle species are grouped into three lineages representing strategies of habitat-type resource partitioning (a split-habitat, migratory pattern, a neritic residence pattern, and a pelagic residence pattern). Within the split-habitat, migratory group, further resource-partitioning by food-type separates the herbivorous Chelonia mydas populations from omnivorous Eretmochelys imbricata and the (apparently) carnivorous Chelonia depressa. Herbivory is seen as integral to the split-habitat, migratory strategy and C. mydas is considered the most “traditional” species, with the migratory habit secondarily lost in the other two. At the same time, the enhanced philopatry selected for by the migration strategy is viewed as responsible for the fact that C. mydas seems to have the most active race-formation of the three species. Further habitat-type partitioning in the neritic group, together with food-type partitioning, separates Caretta caretta from the two Lepidochelys species. L. kempi is represented as a consequence of Panamanian separation from L. olivacea following the last establishment of the isthmus as a land barrier. The third, pelagic residence, strategy is represented by Dermochelys coriacea, with little further differentiation of the line. The paper attempts to show that the evolution of sea turtles has been ecologically logical, that most conceivable niches for marine turtles are presently filled successfully, and that some predictions may be made with regard to gaps in our existing information.