Globally distributed marine taxa are well suited for investigations of biogeographic impacts on genetic diversity, connectivity, and population demography. The sea turtle genus Lepidochelys includes the wide-ranging and abundant olive ridley (L. olivacea), and the geographically restricted and ‘Critically Endangered’ Kemp’s ridley (L. kempii). To investigate their historical biogeography, we analyzed a large dataset of mitochondrial DNA (mtDNA) sequences from olive (n = 943) and Kemp’s (n = 287) ridleys, and genotyped 15 nuclear microsatellite loci in a global sample of olive ridleys (n = 285). We found that the ridley species split ~ 7.5 million years ago, before the Panama Isthmus closure. The most ancient mitochondrial olive ridley lineage, located in the Indian Ocean, was dated to ~ 2.2 Mya. Both mitochondrial and nuclear markers revealed significant structure for olive ridleys between Atlantic (ATL), East Pacific (EP), and Indo-West Pacific (IWP) areas. However, the divergence of mtDNA clades was very recent (< 1 Mya) with low within- clade diversity, supporting a recurrent extinction-recolonization model for these ocean regions. All data showed that ATL and IWP groups were more closely related than those in the EP, with mtDNA data supporting recent recolonization of the ATL from the IWP. Individual olive ridley dispersal between the ATL, EP, and IN/IWP could be interpreted as more male- than female-biased, and genetic diversity was lowest in the Atlantic Ocean. All populations showed signs of recent expansion, and estimated time frames were concordant with their recent colonization history. Investigating species abundance and distribution changes over time is central to evolutionary biology, and this study provides a historical biogeographic context for marine vertebrate conservation and management.