Anopheles mosquitoes transmit the human malaria parasite, a disease that affects hundreds of millions of people globally every year, with Sub-Saharan Africa being the hardest hit. Current control methods, such as indoor residual spraying and insecticide-treated bed nets, are insufficient, prompting the need for new technologies like genetically engineered mosquitoes (GEMs). Once fully developed and profiled in the laboratory, one potential action is to release GEM on a confined environment, such as oceanic islands. In a study analyzing over 20 African islands, the São Tomé and Príncipe Islands, off the coast of Gabon, were identified as the most suitable location for field trials. There is only one vector of malaria on those islands, An. coluzzii, and their populations are genetically isolated from the mainland populations. This study used whole genome sequencing to investigate the spatial dispersal of An. coluzzii within São Tomé and Príncipe Islands. We analyzed data from over 300 specimens collected from various localities on both islands to assess gene flow and population structure. Results showed significant isolation-by-distance for the An. coluzzii population in São Tomé. Conversely, Príncipe presented high inbreeding and no clear population structure or isolation-by-distance. Spatial autocorrelation analysis indicated that individuals within a 7 km radius were more genetically similar to each other. Heterogeneous effective migration rates within each island were identified, highlighting areas of differing resistance or permissiveness to gene flow. These findings are crucial for modelling mosquito dispersal on the islands and provide insights into the potential field performance of GEMs.