Threatened ecosystems such as intertidal mangrove forests often span political boundaries and require cross-border conservation planning initiatives. Population connectivity is key to establish transboundary collaborative actions. In this study, we assessed genetic diversity and connectivity of mangrove populations in a proposed transboundary conservation area (TBCA) between Kenya and Tanzania, where human demography exerts increasing pressure on biodiversity and ecosystem services. East African mangrove ecosystems comprise a complex pattern of estuaries and coastal bays, with Avicennia marina as a major mangrove component. Our main objective was to test a hypothesis of overall migration directionality reflecting regional ocean surface current flows. Fourteen microsatellite markers in 670 A. marina trees revealed no overall significant difference in allele or gene diversity levels between populations but showed an overall effect of geographic distance with a gradient of admixed gene pools. Migration tests and Approximate Bayesian computations supported a customized stepping-stone model of overall south to north migration with bidirectional gene flow and admixture between adjacent bays near the Kenya-Tanzania border. Observed patterns of gene flow suggest an important effect of large rivers and connections via creeks. Our results indicate that relevance for conservation and management of mangroves areas may remain largely within a hydrological connectivity context of each bay, despite prevailing genetic estimates reflecting historically well-connected mangroves between bays. For sustainable management, populations must be considered complementary and not redundant. Hence, transboundary regions must primarily adhere to local contemporary conservation and management, and not solely rely on the assumption of strong regional connectivity built historically.